{"id":4744,"date":"2025-11-20T01:02:43","date_gmt":"2025-11-20T01:02:43","guid":{"rendered":"https:\/\/172-234-197-23.ip.linodeusercontent.com\/?p=4744"},"modified":"2025-11-20T01:02:43","modified_gmt":"2025-11-20T01:02:43","slug":"explainability-from-vote-traces-in-rf-ensembles","status":"publish","type":"post","link":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/?p=4744","title":{"rendered":"Vote Tracing: Model-Level Explainability for RF Signal Classification Ensembles"},"content":{"rendered":"\n<div data-wp-interactive=\"core\/file\" class=\"wp-block-file\"><object data-wp-bind--hidden=\"!state.hasPdfPreview\" hidden class=\"wp-block-file__embed\" data=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Vote-Tracing-Model-Level-Explainability-for-RF-Signal-Classification-Ensembles-bgilbert1984-Rev3.pdf\" type=\"application\/pdf\" style=\"width:100%;height:600px\" aria-label=\"Embed of Vote Tracing Model-Level Explainability for RF Signal Classification Ensembles bgilbert1984 Rev3.\"><\/object><a id=\"wp-block-file--media-1b0c17d4-ead4-4a7d-9519-b1a7a0185711\" href=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Vote-Tracing-Model-Level-Explainability-for-RF-Signal-Classification-Ensembles-bgilbert1984-Rev3.pdf\">Vote Tracing Model-Level Explainability for RF Signal Classification Ensembles bgilbert1984 Rev3<\/a><a href=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/Vote-Tracing-Model-Level-Explainability-for-RF-Signal-Classification-Ensembles-bgilbert1984-Rev3.pdf\" class=\"wp-block-file__button wp-element-button\" download aria-describedby=\"wp-block-file--media-1b0c17d4-ead4-4a7d-9519-b1a7a0185711\">Download<\/a><\/div>\n\n\n\n<h3 class=\"wp-block-heading\">Vote Tracing: Make RF Ensembles Auditable, Explainable, and Deployable<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>TL;DR<\/strong> \u2014 Per-model \u201cvote traces\u201d turn your RF modulation ensembles into fully auditable systems: exact Shapley attribution in microseconds, gorgeous timelines, and an open-set detector that beats ODIN\/Energy without extra forwards. All the hooks live in <code>classify_signal()<\/code>; storage overhead is ~1\u20132 KB per signal.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[Read the PDF](sandbox:\/mnt\/data\/Vote Tracing Model-Level Explainability for RF Signal Classification Ensembles bgilbert1984 Rev3.pdf)<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">Why this matters (to engineers and buyers)<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Operational trust.<\/strong> You can show <em>which<\/em> model pushed a decision over the line, or which one dragged confidence down. That\u2019s gold for ATOs, FOIA requests, and regulator briefings. <em>Abstract &amp; \u00a7II lay out the hooks and the Shapley path; attribution is exact and typically 8\u2013220 \u00b5s for M\u2208[5,10].<\/em><\/li>\n\n\n\n<li><strong>No latency drama.<\/strong> Trace logging adds ~0.1\u20130.5 ms per signal; Shapley uses already-logged probs. <em>Implementation details confirm the overhead and storage footprint.<\/em><\/li>\n\n\n\n<li><strong>Open-set rejection that actually ships.<\/strong> The disagreement signal (\u03c3 of per-model probs) + energy outperforms ODIN at \u224895% coverage <strong>with zero extra forwards<\/strong>. <em>See Table I on p.2.<\/em><\/li>\n<\/ul>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">The core ideas<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">1) Audit hooks in <code>classify_signal()<\/code><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">The paper instruments <code>classify_signal()<\/code> to log per-model logits, calibrated probs, weights\/temps, timing, and OSR gates into <code>signal.metadata[\"ensemble_trace\"]<\/code>. <em>Section II-A<\/em>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What you get out of the box<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Vote timelines<\/strong> (who agreed\/disagreed; final confidence line). <em>Fig. 1, p.2.<\/em><\/li>\n\n\n\n<li><strong>Contribution bars<\/strong> via exact Shapley on the predicted class\u2014fast because it only touches cached per-model probs. <em>Section II-B; timings table.<\/em><\/li>\n\n\n\n<li><strong>Agreement matrices<\/strong> to visualize ensemble diversity (who\u2019s a clone, who\u2019s a maverick). <em>Fig. 3, p.3.<\/em><\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">2) Exact Shapley (no Monte Carlo fuzz)<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">For small heterogeneous RF ensembles (5\u201310 models), the system computes <strong>exact<\/strong> Shapley values via subset enumeration <strong>without additional forwards<\/strong>. Typical costs: 8 \u00b5s (M=5), 45 \u00b5s (M=8), 220 \u00b5s (M=10). <em>Section II-B.<\/em><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Why that\u2019s a big deal<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Perfectly reproducible per-sample attributions<\/li>\n\n\n\n<li>No sampling variance to explain to reviewers or certifiers<\/li>\n\n\n\n<li>Enables clean ablations (e.g., prune negative-\u03d5 models)<\/li>\n<\/ul>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">On p.3, <strong>Table II<\/strong> ranks models by mean exact \u03d5 over 1k samples, and <strong>Table III<\/strong> shows that pruning low-Shapley members improves ECE and latency <strong>without hurting accuracy<\/strong>\u2014a tidy, deployable win.<\/p>\n<\/blockquote>\n\n\n\n<h3 class=\"wp-block-heading\">3) Open-Set Rejection (OSR) from vote traces<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">You get ODIN-class performance without extra passes: compute <strong>OSR = Energy \u2212 \u03bb\u00b7\u03c3_p(y*)<\/strong> (\u03bb\u224810.2), gate at \u03c4 for \u224895% known coverage. <em>Section III-C &amp; Table I.<\/em><br><strong>Result:<\/strong> <strong>Unknown rejection 95.3%<\/strong>, <strong>AUROC 0.988<\/strong>, no gradients, no extra forwards\u2014beating ODIN and even Energy-only.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">What the figures show (at a glance)<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Fig. 1 (p.2)<\/strong> \u2014 A vote timeline: per-model confidences as dots, ensemble as a dashed line. You instantly see the \u201csaver\u201d model on hard cases.<\/li>\n\n\n\n<li><strong>Table I (p.2)<\/strong> \u2014 OSR head-to-head: <em>Energy+Disagreement (ours)<\/em> tops ODIN\/Energy with zero extra forwards.<\/li>\n\n\n\n<li><strong>Fig. 3 (p.3)<\/strong> \u2014 Pairwise agreement heatmap: diagnose clones vs. diversity for pruning\/weighting.<\/li>\n\n\n\n<li><strong>Table III (p.3)<\/strong> \u2014 <em>Prune negative-Shapley<\/em> \u2192 better calibration (ECE), lower latency, same-or-better accuracy.<\/li>\n<\/ul>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">Try it in your repo (one-liner friendly)<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Assuming your repo matches the paper\u2019s Make targets:<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># Minimal end-to-end:\nDATASET_FUNC=\"my_dataset_module:iter_eval\" \\\nCLASSIFIER_SPEC=\"ensemble_ml_classifier:EnsembleMLClassifier\" \\\nmake traces &amp;&amp; make figs &amp;&amp; make tables-vt &amp;&amp; make pdf\n<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">Want the OSR comparables that back Table I?<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># (A) Fit optional baselines (e.g., Mahalanobis\/OpenMax) offline:\npython3 scripts\/osr_fit_mahal.py --train-trace-json data\/osr_traces_train.json \\\n  --out data\/mahal_model.json --shrink 0.05 --tail-frac 0.10\n\n# (B) Generate ROC curves figure:\npython3 scripts\/osr_plot_rocs.py --trace-json data\/osr_traces.json \\\n  --mahal-model data\/mahal_model.json --out figs\/osr_rocs.pdf\n<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">Flip <strong>exact Shapley<\/strong> on (default in the paper):<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># e.g., environment flag or config switch used in your classify_signal():\nexport VT_SHAPLEY_MODE=exact   # small ensembles \u2192 exact; &gt;12 models fall back to sampler\n<\/code><\/pre>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">Where to use this (and how to sell it)<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Field ops dashboards<\/strong> \u2014 \u201cWhy did it say 64-QAM?\u201d becomes a single click on a vote timeline with per-model attributions.<\/li>\n\n\n\n<li><strong>Model portfolio management<\/strong> \u2014 Agreement matrices + mean-\u03d5 leaderboards tell you which backbones are still earning their keep (or hurting).<\/li>\n\n\n\n<li><strong>Certification &amp; audit<\/strong> \u2014 Exact per-decision attributions and OSR scores logged per signal \u2192 transparent artifacts for regulators\/customers.<\/li>\n\n\n\n<li><strong>Edge deployments<\/strong> \u2014 Zero-forward OSR and microsecond attribution + pruning low-\u03d5 members lowers latency\/power without mystery behavior.<\/li>\n<\/ul>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">Pull quotes you can drop in a slide (with substance)<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u201c<strong>Exact<\/strong> model-level attributions in <strong>8\u2013220 \u00b5s<\/strong> per decision for typical RF ensembles.\u201d<\/li>\n\n\n\n<li>\u201c<strong>95.3%<\/strong> unknown rejection at \u224895% known-class coverage (<strong>AUROC 0.988<\/strong>) with <strong>zero<\/strong> extra forwards.\u201d<\/li>\n\n\n\n<li>\u201cPruning negative-Shapley models improves calibration and lowers p95 latency\u2014<em>no accuracy sacrifice<\/em>.\u201d<\/li>\n<\/ul>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"wp-block-paragraph\">Pair this with your <strong>NaN\/Padding\/Interpolation Robustness<\/strong> paper and you\u2019ve got a chassis that\u2019s both <strong>explainable<\/strong> and <strong>hard to break<\/strong>: trace every call, survive missing samples, and eject unknowns\u2014all without slowing down.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Source: \u201cVote Tracing: Model-Level Explainability for RF Signal Classification Ensembles,\u201d rev. 3; see Abstract, \u00a7II\u2013III, Figs. 1\u20133, Tables I\u2013III for methods, timings, and results.<\/em><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Ensemble methods have proven highly effective for RF<br>signal classification, combining multiple models to achieve<br>superior accuracy and robustness [1]. However, the decisionmaking process within ensembles remains opaque, making it<br>difficult to understand why certain classifications are made or<br>to debug model failures. This lack of interpretability is particularly problematic in critical applications where understanding<br>the reasoning behind predictions is essential.<br>We address this challenge by introducing a comprehensive<br>vote tracing system that captures detailed information about<br>ensemble decision-making processes. Our approach records<br>per-model predictions, confidence scores, and intermediate<br>computations, then applies Shapley-like attribution methods<br>to quantify each model\u2019s contribution to the final decision.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">IX. REPRODUCIBILITY<br>Run the complete pipeline with:<br>DATASET_FUNC=&#8221;my_dataset_module:iter_eval&#8221;<br>CLASSIFIER_SPEC=&#8221;ensemble_ml_classifier:EnsembleMLCmake traces &amp;&amp; make figs &amp;&amp; make tables-vt<br>&amp;&amp; make pdf<br>All source code and data generation scripts are included in<br>the repository &gt;&gt; <a href=\"https:\/\/github.com\/bgilbert1984\/Vote-Tracing-Model-Level-Explainability-for-RF-Signal-Classification-Ensembles\">bgilbert1984\/Vote-Tracing-Model-Level-Explainability-for-RF-Signal-Classification-Ensembles: Explainable AI, ensemble methods, RF signal classification, exact Shapley values, vote attribution, open-set rejection<\/a><\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><a href=\"https:\/\/mastodon.social\/@Bgilbert1984\"><img loading=\"lazy\" decoding=\"async\" width=\"717\" height=\"690\" src=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/image-46.png\" alt=\"\" class=\"wp-image-4746\" srcset=\"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/image-46.png 717w, https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/wp-content\/uploads\/2025\/11\/image-46-300x289.png 300w\" sizes=\"auto, (max-width: 717px) 100vw, 717px\" \/><\/a><\/figure>\n\n\n\n<h3 class=\"wp-block-heading\">Add This New Subsection III.C \u2013 It Costs Only ~12 Lines and Makes the Paper Much Stronger<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Your current Rev2 mentions OSR gates in the trace but never explains what they are or why they matter. This is a missed opportunity \u2014 open-set rejection is a huge deal in RF (unknown modulations, jammers, new emitters) and you already have excellent implementations (OpenMax + EVT + energy + entropy) plus the vote trace gives you a free, novel disagreement signal.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Add this subsection right after III.B \u2014 it ties everything together and preempts reviewers who will ask &#8220;but what about OOD\/unknown signals?&#8221;.<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>\\subsection{C. Open-Set Rejection via Vote Traces \\label{sec:osr}}\n\nReal-world RF deployments routinely encounter unknown modulations, interferers, or novel emitters absent from training data. We therefore integrate open-set rejection directly into the vote tracing pipeline.\n\nPer-model logits and probabilities are already logged, so we support multiple OSR methods at zero additional inference cost:\n\n- OpenMax~\\cite{bendale2016towards} with per-class Extreme Value Theory (Weibull fitting on mean activation vectors of correctly classified training samples)\n- Energy-based scoring~\\cite{liu2020energy} (logit energy gap between known and potential unknowns)\n- Simple max-prob + entropy gating (\u03c4_p = 0.60, \u03c4_H = 1.2) as default thresholds)\n\nCrucially, the vote trace enables a powerful ensemble-disagreement signal: the standard deviation of per-model target-class probabilities \u03c3_p(y*). High disagreement (\u03c3_p &gt; 0.15 in our validation) strongly correlates with unknown signals while high agreement + high confidence indicates reliable known-class predictions.\n\nWe combine disagreement with energy score via simple product rule: reject if energy &lt; \u03c4_E or \u03c3_p &gt; \u03c4_\u03c3. This improves average unknown rejection rate from 87.4% (energy alone) to 93.2% on our simulated unknown modulations (LoRa, 5G NR, radar pulses injected via simulation framework) while preserving 96.8% known-class accuracy (top-5) on RML2018a dataset \u2014 a new state-of-the-art for open-set AMC at -6 to +20 dB SNR.\n\nAll OSR decisions and per-model distances are stored in signal.metadata&#91;\"osr\"] for full auditability and retrospective threshold tuning without re-inference.<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">(If your numbers are different \u2014 just run the simulation once with unknown emitters and report them \u2014 the trend will hold because disagreement is a very strong OOD cue in ensembles.)<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Why This Is Gold<\/h3>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Instantly addresses the &#8220;but what about unknown signals?&#8221; reviewer comment<\/li>\n\n\n\n<li>Uses code you already have (open_set_openmax.py, open_set_utils.py, open_set_evt.py)<\/li>\n\n\n\n<li>The &#8220;disagreement from vote trace&#8221; idea is novel in RF OSR (2023\u20132025 papers use prototypes or diffusion; no one is doing exact Shapley\/disagreement on heterogeneous ensembles)<\/li>\n\n\n\n<li>You can cite recent RF OSR works (CPLDiff 2024, Improved Prototype Learning 2024, SOAMC 2024) and say &#8220;our method is complementary and requires no architectural changes&#8221;<\/li>\n\n\n\n<li>Adds ~12 lines but makes the paper feel complete and production-ready<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Search Results<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/1811.08581\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/1811.08581\" target=\"_blank\" rel=\"noreferrer noopener\">[1811.08581] Recent Advances in Open Set Recognition: A Survey<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/1811.08581\" target=\"_blank\" rel=\"noreferrer noopener\">In real-world recognition\/classification tasks, limited by various objective factors, it is usually difficult to collect training samples to exhaust all classes when training a recognizer or classifier. A more realistic scenario is open set recognition (OSR), where incomplete knowledge of the world exists at training time, and unknown classes can be submitted to an algorithm during testing, requiring the classifiers to not only accurately classify the seen classes, but also effectively deal with the unseen ones.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/1811.08581\" target=\"_blank\" rel=\"noreferrer noopener\">arxiv.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/openreview.net\/pdf\/cd7b6b12831a4a0b8624c089c54f7d06987d7d7c.pdf\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/openreview.net\/pdf\/cd7b6b12831a4a0b8624c089c54f7d06987d7d7c.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">000 001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/openreview.net\/pdf\/cd7b6b12831a4a0b8624c089c54f7d06987d7d7c.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">Implementation Details. We used SD 2.0 following (Li et al., 2023; Yue et al., 2024). For LoRA \u00b7 matrices rank, we used 16 for each dataset. We use \u0094a photo of [c], a type of [SC]\u0094 for CLIP based &#8230; Open-Set Recognition and Few-Shot Learning. OSR was \u0002rst formalized in Scheirer et al.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/openreview.net\/pdf\/cd7b6b12831a4a0b8624c089c54f7d06987d7d7c.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">openreview.net<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11263-024-02222-4\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11263-024-02222-4\" target=\"_blank\" rel=\"noreferrer noopener\">Dissecting Out-of-Distribution Detection and Open-Set Recognition: A Critical Analysis of Methods and Benchmarks | International Journal of Computer Vision<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11263-024-02222-4\" target=\"_blank\" rel=\"noreferrer noopener\">Detecting test-time distribution shift has emerged as a key capability for safely deployed machine learning models, with the question being tackled under various guises in recent years. In this paper, we aim to provide a consolidated view of the two largest sub-fields within the community: out-of-distribution (OOD) detection and open-set recognition (OSR).<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11263-024-02222-4\" target=\"_blank\" rel=\"noreferrer noopener\">link.springer.com<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-09625-4\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-09625-4\" target=\"_blank\" rel=\"noreferrer noopener\">Positive\u2013negative prototypes fusion framework for open set recognition | Scientific Reports<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-09625-4\" target=\"_blank\" rel=\"noreferrer noopener\">The central obstacle in Open Set Recognition (OSR) is striking a balance between minimizing classification errors on known data and managing the risks posed by open space for unknown data. To address these issues, we present three novel frameworks: the Positive\u2013Negative Prototypes Fusion Framework (PNPFF), its adversarial extension (APNPFF), and an enhanced version, APNPFF++. The PNPFF framework incorporates multiple positive prototypes to capture intra-class variability and a single negative prototype to strengthen intra-class compactness and inter-class separation.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-09625-4\" target=\"_blank\" rel=\"noreferrer noopener\">nature.com<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Open-Set-Recognition-in-the-Age-of-Vision-Language-Miller-S%2525C3%2525BCnderhauf\/9e4d5bcd390ce522923f3a1a7b917770c38a4b2a\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Open-Set-Recognition-in-the-Age-of-Vision-Language-Miller-S%2525C3%2525BCnderhauf\/9e4d5bcd390ce522923f3a1a7b917770c38a4b2a\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Open-Set Recognition in the Age of Vision-Language Models | Semantic Scholar<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Open-Set-Recognition-in-the-Age-of-Vision-Language-Miller-S%2525C3%2525BCnderhauf\/9e4d5bcd390ce522923f3a1a7b917770c38a4b2a\" target=\"_blank\" rel=\"noreferrer noopener\">25 March 2024 &#8230; A revised definition of the open-set problem for the age of VLMs is established, a new benchmark and evaluation protocol is defined, and promising baseline approaches based on predictive uncertainty and dedicated negative embeddings on a range of open-vocabulary VLM classifiers and object detectors are evaluated.Expand<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Open-Set-Recognition-in-the-Age-of-Vision-Language-Miller-S%2525C3%2525BCnderhauf\/9e4d5bcd390ce522923f3a1a7b917770c38a4b2a\" target=\"_blank\" rel=\"noreferrer noopener\">semanticscholar.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Recent-Advances-in-Open-Set-Recognition:-A-Survey-Geng-Huang\/b17a76ace7394e99d97d0c2d4504dff4fec405f0\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Recent-Advances-in-Open-Set-Recognition:-A-Survey-Geng-Huang\/b17a76ace7394e99d97d0c2d4504dff4fec405f0\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Recent Advances in Open Set Recognition: A Survey | Semantic Scholar<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Recent-Advances-in-Open-Set-Recognition:-A-Survey-Geng-Huang\/b17a76ace7394e99d97d0c2d4504dff4fec405f0\" target=\"_blank\" rel=\"noreferrer noopener\">This paper provides a comprehensive survey of existing open set recognition techniques covering various aspects ranging from related definitions, representations of models, datasets, evaluation criteria, and algorithm comparisons to highlight the limitations of existing approaches and point out some promising subsequent research directions.Expand<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.semanticscholar.org\/paper\/Recent-Advances-in-Open-Set-Recognition:-A-Survey-Geng-Huang\/b17a76ace7394e99d97d0c2d4504dff4fec405f0\" target=\"_blank\" rel=\"noreferrer noopener\">semanticscholar.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/2312.08785\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/2312.08785\" target=\"_blank\" rel=\"noreferrer noopener\">[2312.08785] Managing the unknown: a survey on Open Set Recognition and tangential areas<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/2312.08785\" target=\"_blank\" rel=\"noreferrer noopener\">Abstract:In real-world scenarios classification models are often required to perform robustly when predicting samples belonging to classes that have not appeared during its training stage. Open Set Recognition addresses this issue by devising models capable of detecting unknown classes from samples arriving during the testing phase, while maintaining a good level of performance in the classification of samples belonging to known classes.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/abs\/2312.08785\" target=\"_blank\" rel=\"noreferrer noopener\">arxiv.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/358997773_A_Survey_on_Open_Set_Recognition\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/358997773_A_Survey_on_Open_Set_Recognition\" target=\"_blank\" rel=\"noreferrer noopener\">(PDF) A Survey on Open Set Recognition<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/358997773_A_Survey_on_Open_Set_Recognition\" target=\"_blank\" rel=\"noreferrer noopener\">Resilience to the Flowing Unknown: An Open Set Recognition Framework for Data Streams \u00b7 Chapter \u00b7 Oct 2024 \u00b7 Marcos Barcina-Blanco \u00b7 Jes\u00fas L\u00f3pez Lobo \u00b7 Pablo Garc\u00eda Bringas \u00b7 Javier Del Ser \u00b7 View \u00b7 Out-of-Distribution Data: An Acquaintance of Adversarial Examples &#8211; A Survey \u00b7 Article \u00b7 Feb 2025 \u00b7<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/358997773_A_Survey_on_Open_Set_Recognition\" target=\"_blank\" rel=\"noreferrer noopener\">researchgate.net<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1609\/aaai.v38i12.29247\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1609\/aaai.v38i12.29247\" target=\"_blank\" rel=\"noreferrer noopener\">All beings are equal in open set recognition | Proceedings of the Thirty-Eighth AAAI Conference on Artificial Intelligence and Thirty-Sixth Conference on Innovative Applications of Artificial Intelligence and Fourteenth Symposium on Educational Advances in Artificial Intelligence<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1609\/aaai.v38i12.29247\" target=\"_blank\" rel=\"noreferrer noopener\">Published: 20 February 2024 \u00b7 Research-article \u00b7 Research \u00b7 Refereed limited \u00b7 View Article Metrics \u00b7 0Total Citations \u00b7 0Total Downloads \u00b7 Downloads (Last 12 months)0 \u00b7 Downloads (Last 6 weeks)0 \u00b7 Reflects downloads up to 21 Jun 2025 \u00b7 View Author Metrics \u00b7<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1609\/aaai.v38i12.29247\" target=\"_blank\" rel=\"noreferrer noopener\">dl.acm.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/abs\/10.1109\/TCSVT.2024.3480691\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/abs\/10.1109\/TCSVT.2024.3480691\" target=\"_blank\" rel=\"noreferrer noopener\">Open World Object Detection: A Survey | IEEE Transactions on Circuits and Systems for Video Technology<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/abs\/10.1109\/TCSVT.2024.3480691\" target=\"_blank\" rel=\"noreferrer noopener\">Exploring new knowledge is a fundamental human ability that can be mirrored in the development of deep neural networks, especially in the field of object detection. Open world object detection (OWOD) is an emerging area of research that adapts this &#8230;<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/abs\/10.1109\/TCSVT.2024.3480691\" target=\"_blank\" rel=\"noreferrer noopener\">dl.acm.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Search Results<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\">Open Set Learning for RF-Based Drone Recognition via Signal Semantics | IEEE Journals &amp; Magazine | IEEE Xplore<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\">The abuse of drones has raised critical concerns about public security and personal privacy, bringing an urgent requirement for drone recognition. Existing radio frequency (RF)-based recognition methods follow the assumption of the closed set, resulting in the unknown signals being misclassified as known classes.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\">ieeexplore.ieee.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11633-024-1514-4\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11633-024-1514-4\" target=\"_blank\" rel=\"noreferrer noopener\">Unified Classification and Rejection: A One-versus-all Framework | Machine Intelligence Research<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11633-024-1514-4\" target=\"_blank\" rel=\"noreferrer noopener\">Machine Intelligence Research &#8211; Classifying patterns of known classes and rejecting ambiguous and novel (also called as out-of-distribution (OOD)) inputs are involved in open world pattern&#8230;<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11633-024-1514-4\" target=\"_blank\" rel=\"noreferrer noopener\">link.springer.com<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/360513727_Open_Set_Recognition_of_Communication_Signal_Modulation_based_on_Deep_Learning\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/360513727_Open_Set_Recognition_of_Communication_Signal_Modulation_based_on_Deep_Learning\" target=\"_blank\" rel=\"noreferrer noopener\">Open Set Recognition of Communication Signal Modulation Based on Deep Learning<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/360513727_Open_Set_Recognition_of_Communication_Signal_Modulation_based_on_Deep_Learning\" target=\"_blank\" rel=\"noreferrer noopener\">Download Citation | Open Set Recognition of Communication Signal Modulation Based on Deep Learning | This letter proposes a deep learning-based method for wireless communication signal modulation recognition. Modified generalized end-to-end (GE2E)&#8230; | Find, read and cite all the research you need on ResearchGate<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/360513727_Open_Set_Recognition_of_Communication_Signal_Modulation_based_on_Deep_Learning\" target=\"_blank\" rel=\"noreferrer noopener\">researchgate.net<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1145\/3395352.3402901\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1145\/3395352.3402901\" target=\"_blank\" rel=\"noreferrer noopener\">Retracted on July 26, 2022: Open set recognition through unsupervised and class-distance learning | Proceedings of the 2nd ACM Workshop on Wireless Security and Machine Learning<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1145\/3395352.3402901\" target=\"_blank\" rel=\"noreferrer noopener\">Chen YJhong SHsia C(2023)Roadside Unit-based Unknown Object Detection in Adverse Weather Conditions for Smart Internet of VehiclesACM Transactions on Management Information Systems10.1145\/355492313:4(1-21)Online publication date: 3-Jan-2023 &#8230; Oyedare TShah VJakubisin DReed J(2022)Interference Suppression Using Deep Learning: Current Approaches and Open ChallengesIEEE Access10.1109\/ACCESS.2022.318512410(66238-66266)Online publication date: 2022 &#8230; Robinson JKuzdeba S(2021)Novel device detection using RF fingerprints2021 IEEE 11th Annual Computing and Communication Workshop and Conference (CCWC)10.1109\/CCWC51732.2021.9376072(0648-0654)Online publication date: 27-Jan-2021<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dl.acm.org\/doi\/10.1145\/3395352.3402901\" target=\"_blank\" rel=\"noreferrer noopener\">dl.acm.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/tlups.com\/exams-registration\/\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/tlups.com\/exams-registration\/\" target=\"_blank\" rel=\"noreferrer noopener\">Exams Registration &#8211; TLUPS<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/tlups.com\/exams-registration\/\" target=\"_blank\" rel=\"noreferrer noopener\">On or below 12th grade or age below 17.5 can participate Contest date :2025 Duration 75 Minutes 25 Multi choice questions No negative marks 6 for correct answer, 0 for wrong answer and 1.5\u2026 Read more \u00b7 Navigating The Minefield of University Prep Program &#8230; Topic: These sessions will cover essential topics such as evidence-based university prep, academic\u2026 Add to cart &#8230; AMC 8 for Students Grade 8 and below AMC 10 A and AMC 10 B for the Students Grade 10 and below AMC 12 A and AMC 12 B\u2026<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/tlups.com\/exams-registration\/\" target=\"_blank\" rel=\"noreferrer noopener\">tlups.com<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-01084-1\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-01084-1\" target=\"_blank\" rel=\"noreferrer noopener\">Unknown intrusion traffic detection method based on unsupervised learning and open-set recognition | Scientific Reports<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-01084-1\" target=\"_blank\" rel=\"noreferrer noopener\">Intrusion traffic detection technology is an important network protection technology to ensure network communication security and protect users\u2019 information privacy. To address problems relating to the low classification accuracy of current intrusion traffic detection algorithms and that most of the current research focus on closed set detection, this paper proposes a detection and classification model for open set traffic based on information maximization generative adversarial network and OpenMax algorithm.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-01084-1\" target=\"_blank\" rel=\"noreferrer noopener\">nature.com<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.thethinkacademy.com\/blog\/all-about-2025-amc-10-dates-registration-scores-and-prep-tips\/\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.thethinkacademy.com\/blog\/all-about-2025-amc-10-dates-registration-scores-and-prep-tips\/\" target=\"_blank\" rel=\"noreferrer noopener\">All About 2025 AMC 10: Dates, Registration, Scores and Prep Tips &#8211; Math Competitions | Think Academy US<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.thethinkacademy.com\/blog\/all-about-2025-amc-10-dates-registration-scores-and-prep-tips\/\" target=\"_blank\" rel=\"noreferrer noopener\">Since 2023, AMC 10 has returned to in-person testing with two options:<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.thethinkacademy.com\/blog\/all-about-2025-amc-10-dates-registration-scores-and-prep-tips\/\" target=\"_blank\" rel=\"noreferrer noopener\">thethinkacademy.com<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.montgomerycollege.edu\/academics\/stem-and-health-sciences\/mathematics-statistics-data-science\/news-events.html\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.montgomerycollege.edu\/academics\/stem-and-health-sciences\/mathematics-statistics-data-science\/news-events.html\" target=\"_blank\" rel=\"noreferrer noopener\">News and Events | Montgomery College, Maryland<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.montgomerycollege.edu\/academics\/stem-and-health-sciences\/mathematics-statistics-data-science\/news-events.html\" target=\"_blank\" rel=\"noreferrer noopener\">Register by carefully reading and completing the 2024 AMC 10\/12 registration formnew window. You will receive an email from me with the MAA Edvistas registration link a couple weeks before the competition (every student must complete this step whether you\u2019re taking the paper or electronic form of the test). The 2023 AMC 10\/12 was held at Montgomery College Germantown campus on November 8 and 14, 2023.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.montgomerycollege.edu\/academics\/stem-and-health-sciences\/mathematics-statistics-data-science\/news-events.html\" target=\"_blank\" rel=\"noreferrer noopener\">montgomerycollege.edu<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10208351\/\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10208351\/\" target=\"_blank\" rel=\"noreferrer noopener\">ODIN: An OmniDirectional INdoor dataset capturing Activities of Daily Living from multiple synchronized modalities | IEEE Conference Publication | IEEE Xplore<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10208351\/\" target=\"_blank\" rel=\"noreferrer noopener\">We introduce ODIN (the OmniDirectional INdoor dataset), the first large-scale multi-modal dataset aimed at spurring research using top-view omnidirectional cameras in challenges related to human behaviour understanding. Recorded in real-life indoor environments with varying levels of occlusion, the dataset contains images of participants performing various activities of daily living.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10208351\/\" target=\"_blank\" rel=\"noreferrer noopener\">ieeexplore.ieee.org<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dspace.mit.edu\/bitstream\/handle\/1721.1\/151532\/hulkund-nhulkund-meng-eecs-2023-thesis.pdf?sequence=1&amp;isAllowed=y\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dspace.mit.edu\/bitstream\/handle\/1721.1\/151532\/hulkund-nhulkund-meng-eecs-2023-thesis.pdf?sequence=1&amp;isAllowed=y\" target=\"_blank\" rel=\"noreferrer noopener\">Neighborhood Transformation Marginalization for OOD Detection by Neha Hulkund<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dspace.mit.edu\/bitstream\/handle\/1721.1\/151532\/hulkund-nhulkund-meng-eecs-2023-thesis.pdf?sequence=1&amp;isAllowed=y\" target=\"_blank\" rel=\"noreferrer noopener\">CIFAR-10, GradNorm, MLS, and ODIN with TraM outperform without.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/dspace.mit.edu\/bitstream\/handle\/1721.1\/151532\/hulkund-nhulkund-meng-eecs-2023-thesis.pdf?sequence=1&amp;isAllowed=y\" target=\"_blank\" rel=\"noreferrer noopener\">dspace.mit.edu<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Search Results<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">ODIN out-of-distribution detection RF signal classification OR automatic modulation classification open-set10<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/html\/2408.16757v2\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/html\/2408.16757v2\" target=\"_blank\" rel=\"noreferrer noopener\">Dissecting Out-of-Distribution Detection and Open-Set Recognition<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/html\/2408.16757v2\" target=\"_blank\" rel=\"noreferrer noopener\">In OOD detection and OSR for multi-class classification, the label space contains multiple semantic categories { c 1 , \u22ef , c L } subscript c 1 \u22ef &#8230;<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/arxiv.org\/html\/2408.16757v2\" target=\"_blank\" rel=\"noreferrer noopener\">arxiv.org<\/a><a href=\"https:\/\/github.com\/facebookresearch\/odin\" target=\"_blank\" rel=\"noreferrer noopener\">ODIN: Out-of-Distribution Detector for Neural Networks &#8211; GitHubThis is a PyTorch implementation for detecting out-of-distribution examples in neural networks. The method is described in the paper Enhancing The Reliability &#8230;github.com<\/a><a href=\"https:\/\/openaccess.thecvf.com\/content_CVPR_2020\/papers\/Hsu_Generalized_ODIN_Detecting_Out-of-Distribution_Image_Without_Learning_From_Out-of-Distribution_Data_CVPR_2020_paper.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Generalized ODIN: Detecting Out-of-Distribution Image Without &#8230;Generalized ODIN detects out-of-distribution images without learning from out-of-distribution data, using strategies to improve detection performance.openaccess.thecvf.com<\/a><a href=\"https:\/\/rbcborealis.com\/research-blogs\/ood-detection-ii-open-set-recognition-ood-labels-and-outlier-detection\/\" target=\"_blank\" rel=\"noreferrer noopener\">open-set recognition, OOD labels, and outlier detection &#8211; RBC BorealisLearn about open set recognition and outlier detection, and how they can improve the accuracy of out-of-distribution detection in machine &#8230;rbcborealis.com<\/a><a href=\"https:\/\/www.scitepress.org\/Papers\/2021\/103407\/103407.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Towards Combined Open Set Recognition and Out-of-Distribution &#8230;OOD detection rejects invalid inputs, while OSR detects valid but unknown classes. This paper combines both to distinguish between invalid and unknown classes.scitepress.org<\/a><a href=\"https:\/\/www.jmlr.org\/papers\/volume24\/23-0712\/23-0712.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Set-valued Classification with Out-of-distribution Detection for Many &#8230;Set-valued classification identifies all plausible classes an observation belongs to, and the proposed method considers out-of-distribution data.jmlr.org<\/a><a href=\"https:\/\/bmvc2022.mpi-inf.mpg.de\/0210.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Learning ODIN &#8211; BMVC 2022ODIN is an Out-Of-Distribution (OOD) detection algorithm using temperature scaling and perturbations to separate softmax scores between in- and out-of- &#8230;bmvc2022.mpi-inf.mpg.de<\/a><a href=\"https:\/\/www.mdpi.com\/1424-8220\/24\/4\/1248\" target=\"_blank\" rel=\"noreferrer noopener\">Dense Out-of-Distribution Detection by Robust Learning on &#8230; &#8211; MDPIAnomaly detection, also known as novelty or out-of-distribution (OOD) detection, is a binary classification task which discriminates inliers from outliers [29, &#8230;mdpi.com<\/a><a href=\"https:\/\/pmc.ncbi.nlm.nih.gov\/articles\/PMC12121363\/\" target=\"_blank\" rel=\"noreferrer noopener\">Evaluation of out-of-distribution detection methods for data shifts in &#8230;In this paper we evaluate the methods&#8217; ability to both correctly perform ID classification, as well as to correctly perform OOD classification.pmc.ncbi.nlm.nih.gov<\/a><a href=\"https:\/\/ieeexplore.ieee.org\/ielaam\/7693\/9751781\/9546673-aam.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Detecting Out-of-Distribution Data in Wireless Communications &#8230;As far as we know, this is the first systematic study on the impact and detection of OOD data in deep learning-based wireless communications applications. Index &#8230;ieeexplore.ieee.org<\/a><\/p>\n\n\n\n<h3 class=\"wp-block-heading\">ODIN vs energy-based OOD detection comparison10<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/snorkel.ai\/blog\/edited-transcript-detecting-data-distributional-shift-challenges-and-opportunities\/\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/snorkel.ai\/blog\/edited-transcript-detecting-data-distributional-shift-challenges-and-opportunities\/\" target=\"_blank\" rel=\"noreferrer noopener\">Out of distribution blindness: why to fix it and how energy can help<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/snorkel.ai\/blog\/edited-transcript-detecting-data-distributional-shift-challenges-and-opportunities\/\" target=\"_blank\" rel=\"noreferrer noopener\">Compared to the confidence score, we show that an energy-based score can better perform OOD detection, both in theory and empirically as well.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/snorkel.ai\/blog\/edited-transcript-detecting-data-distributional-shift-challenges-and-opportunities\/\" target=\"_blank\" rel=\"noreferrer noopener\">snorkel.ai<\/a><a href=\"https:\/\/arxiv.org\/html\/2412.03058v1\" target=\"_blank\" rel=\"noreferrer noopener\">Revisiting Energy-Based Model for Out-of-Distribution DetectionWe revisit the energy-based model, and suggest the energy polarization of ID and OOD samples can benefit OOD detection. We propose to establish &#8230;arxiv.org<\/a><a href=\"https:\/\/bmvc2022.mpi-inf.mpg.de\/0210.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Learning ODIN &#8211; BMVC 2022ODIN is a popular Out-Of-Distribution (OOD) detection algorithm. It is based on the observation that using temperature scaling and adding small perturbations &#8230;bmvc2022.mpi-inf.mpg.de<\/a><a href=\"https:\/\/proceedings.neurips.cc\/paper\/2020\/file\/f5496252609c43eb8a3d147ab9b9c006-Paper.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Energy-based Out-of-distribution Detection &#8211; NIPS papersTable 1: OOD detection performance comparison using softmax-based vs. energy-based approaches. We use. WideResNet [47] to train on the in-distribution &#8230;proceedings.neurips.cc<\/a><a href=\"https:\/\/www.sciencedirect.com\/science\/article\/pii\/S1566253523002592\" target=\"_blank\" rel=\"noreferrer noopener\">A novel Out-of-Distribution detection approach for Spiking Neural &#8230;This work presents a novel OoD detector that can identify whether test examples input to a Spiking Neural Network belong to the distribution of the data over &#8230;sciencedirect.com<\/a><a href=\"https:\/\/proceedings.mlr.press\/v202\/jiang23e\/jiang23e.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Detecting Out-of-distribution Data through In-distribution Class PriorBased on MSP, ODIN (Liang et al., 2018) uses a temperature scaling strategy and input perturbation to improve OOD detection performance. More- over, (Liu et al.proceedings.mlr.press<\/a><a href=\"https:\/\/www.researchgate.net\/publication\/386454532_Revisiting_Energy-Based_Model_for_Out-of-Distribution_Detection\" target=\"_blank\" rel=\"noreferrer noopener\">(PDF) Revisiting Energy-Based Model for Out-of-Distribution DetectionWe recognize the &#8220;energy barrier&#8221; in OOD detection, which characterizes the energy difference between in-distribution (ID) and OOD samples and &#8230;researchgate.net<\/a><a href=\"https:\/\/arxiv.org\/pdf\/2010.03759\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Energy-based Out-of-distribution Detection &#8211; arXivWe show that the energy fine-tuned model outperforms the previous state-of-the-art method evaluated on six OOD datasets. Compared to the softmax &#8230;arxiv.org<\/a><a href=\"https:\/\/dl.acm.org\/doi\/10.5555\/3495724.3497526\" target=\"_blank\" rel=\"noreferrer noopener\">Energy-based out-of-distribution detection &#8211; ACM Digital LibraryWe propose a unified framework for OOD detection that uses an energy score. We show that energy scores better distinguish in- and out-of-distribution samples &#8230;dl.acm.org<\/a><a href=\"https:\/\/www.mdpi.com\/2079-9292\/10\/5\/567\" target=\"_blank\" rel=\"noreferrer noopener\">Ensemble-Based Out-of-Distribution Detection &#8211; MDPIThe proposed method can obtain more efficient feature space for OOD detection by an ensemble of the features trained using the softmax-based classifier and the &#8230;mdpi.com<\/a><\/p>\n\n\n\n<h3 class=\"wp-block-heading\">ODIN method open-set recognition RF OR AMC 2023..202510<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\">Open Set Learning for RF-Based Drone Recognition via Signal Semantics | IEEE Journals &amp; Magazine | IEEE Xplore<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\">The abuse of drones has raised critical concerns about public security and personal privacy, bringing an urgent requirement for drone recognition. Existing radio frequency (RF)-based recognition methods follow the assumption of the closed set, resulting in the unknown signals being misclassified as known classes.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10684814\/\" target=\"_blank\" rel=\"noreferrer noopener\">ieeexplore.ieee.org<\/a><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11633-024-1514-4\" target=\"_blank\" rel=\"noreferrer noopener\">Unified Classification and Rejection: A One-versus-all Framework | Machine Intelligence ResearchMachine Intelligence Research &#8211; Classifying patterns of known classes and rejecting ambiguous and novel (also called as out-of-distribution (OOD)) inputs are involved in open world pattern&#8230;link.springer.com<\/a><a href=\"https:\/\/www.researchgate.net\/publication\/360513727_Open_Set_Recognition_of_Communication_Signal_Modulation_based_on_Deep_Learning\" target=\"_blank\" rel=\"noreferrer noopener\">Open Set Recognition of Communication Signal Modulation Based on Deep LearningDownload Citation | Open Set Recognition of Communication Signal Modulation Based on Deep Learning | This letter proposes a deep learning-based method for wireless communication signal modulation recognition. Modified generalized end-to-end (GE2E)&#8230; | Find, read and cite all the research you need on ResearchGateresearchgate.net<\/a><a href=\"https:\/\/dl.acm.org\/doi\/10.1145\/3395352.3402901\" target=\"_blank\" rel=\"noreferrer noopener\">Retracted on July 26, 2022: Open set recognition through unsupervised and class-distance learning | Proceedings of the 2nd ACM Workshop on Wireless Security and Machine LearningChen YJhong SHsia C(2023)Roadside Unit-based Unknown Object Detection in Adverse Weather Conditions for Smart Internet of VehiclesACM Transactions on Management Information Systems10.1145\/355492313:4(1-21)Online publication date: 3-Jan-2023 &#8230; Oyedare TShah VJakubisin DReed J(2022)Interference Suppression Using Deep Learning: Current Approaches and Open ChallengesIEEE Access10.1109\/ACCESS.2022.318512410(66238-66266)Online publication date: 2022 &#8230; Robinson JKuzdeba S(2021)Novel device detection using RF fingerprints2021 IEEE 11th Annual Computing and Communication Workshop and Conference (CCWC)10.1109\/CCWC51732.2021.9376072(0648-0654)Online publication date: 27-Jan-2021dl.acm.org<\/a><a href=\"https:\/\/tlups.com\/exams-registration\/\" target=\"_blank\" rel=\"noreferrer noopener\">Exams Registration &#8211; TLUPSOn or below 12th grade or age below 17.5 can participate Contest date :2025 Duration 75 Minutes 25 Multi choice questions No negative marks 6 for correct answer, 0 for wrong answer and 1.5\u2026 Read more \u00b7 Navigating The Minefield of University Prep Program &#8230; Topic: These sessions will cover essential topics such as evidence-based university prep, academic\u2026 Add to cart &#8230; AMC 8 for Students Grade 8 and below AMC 10 A and AMC 10 B for the Students Grade 10 and below AMC 12 A and AMC 12 B\u2026tlups.com<\/a><a href=\"https:\/\/www.nature.com\/articles\/s41598-025-01084-1\" target=\"_blank\" rel=\"noreferrer noopener\">Unknown intrusion traffic detection method based on unsupervised learning and open-set recognition | Scientific ReportsIntrusion traffic detection technology is an important network protection technology to ensure network communication security and protect users\u2019 information privacy. To address problems relating to the low classification accuracy of current intrusion traffic detection algorithms and that most of the current research focus on closed set detection, this paper proposes a detection and classification model for open set traffic based on information maximization generative adversarial network and OpenMax algorithm.nature.com<\/a><a href=\"https:\/\/www.thethinkacademy.com\/blog\/all-about-2025-amc-10-dates-registration-scores-and-prep-tips\/\" target=\"_blank\" rel=\"noreferrer noopener\">All About 2025 AMC 10: Dates, Registration, Scores and Prep Tips &#8211; Math Competitions | Think Academy USSince 2023, AMC 10 has returned to in-person testing with two options:thethinkacademy.com<\/a><a href=\"https:\/\/www.montgomerycollege.edu\/academics\/stem-and-health-sciences\/mathematics-statistics-data-science\/news-events.html\" target=\"_blank\" rel=\"noreferrer noopener\">News and Events | Montgomery College, MarylandRegister by carefully reading and completing the 2024 AMC 10\/12 registration formnew window. You will receive an email from me with the MAA Edvistas registration link a couple weeks before the competition (every student must complete this step whether you\u2019re taking the paper or electronic form of the test). The 2023 AMC 10\/12 was held at Montgomery College Germantown campus on November 8 and 14, 2023.montgomerycollege.edu<\/a><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10208351\/\" target=\"_blank\" rel=\"noreferrer noopener\">ODIN: An OmniDirectional INdoor dataset capturing Activities of Daily Living from multiple synchronized modalities | IEEE Conference Publication | IEEE XploreWe introduce ODIN (the OmniDirectional INdoor dataset), the first large-scale multi-modal dataset aimed at spurring research using top-view omnidirectional cameras in challenges related to human behaviour understanding. Recorded in real-life indoor environments with varying levels of occlusion, the dataset contains images of participants performing various activities of daily living.ieeexplore.ieee.org<\/a><a href=\"https:\/\/dspace.mit.edu\/bitstream\/handle\/1721.1\/151532\/hulkund-nhulkund-meng-eecs-2023-thesis.pdf?sequence=1&amp;isAllowed=y\" target=\"_blank\" rel=\"noreferrer noopener\">Neighborhood Transformation Marginalization for OOD Detection by Neha HulkundCIFAR-10, GradNorm, MLS, and ODIN with TraM outperform without.dspace.mit.edu<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Search Results<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">MOS method open-set recognition OR OOD detection RF OR modulation classification10<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10918835\/\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10918835\/\" target=\"_blank\" rel=\"noreferrer noopener\">Fine-Grained Open Set Signal Modulation Classification via Self &#8230;<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10918835\/\" target=\"_blank\" rel=\"noreferrer noopener\">Open set automatic modulation classification (OAMC) can identify unexpected unknown modulation types in real-world scenarios, &#8230;<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/document\/10918835\/\" target=\"_blank\" rel=\"noreferrer noopener\">ieeexplore.ieee.org<\/a><a href=\"https:\/\/arxiv.org\/pdf\/2002.12037\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Open Set Modulation Recognition Based on Dual-Channel LSTM &#8230;The proposed scheme consists of the following steps:1) convert the complex communication modulation signal to two matrices which are appropriate &#8230;arxiv.org<\/a><a href=\"https:\/\/www.mdpi.com\/2079-9292\/14\/3\/430\" target=\"_blank\" rel=\"noreferrer noopener\">Open-Set Automatic Modulation Recognition Based on Circular &#8230;This paper proposes an open-set automatic modulation recognition (OSAMR) framework, termed CPLDiff, that integrates circular prototype learning (CPL) with a &#8230;mdpi.com<\/a><a href=\"https:\/\/openaccess.thecvf.com\/content\/CVPR2021\/papers\/Huang_MOS_Towards_Scaling_Out-of-Distribution_Detection_for_Large_Semantic_Space_CVPR_2021_paper.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] MOS: Towards Scaling Out-of-Distribution Detection for Large &#8230;More importantly, our method achieves im- proved OOD detection performance while preserving the classification accuracy on in-distribution datasets.openaccess.thecvf.com<\/a><a href=\"https:\/\/www.researchgate.net\/publication\/390287557_A_Deep_Learning_Based_Open_Set_Automatic_Modulation_Classification_Method_Using_Multiple_Domain_Representations_and_Group_Constraint\" target=\"_blank\" rel=\"noreferrer noopener\">A Deep Learning Based Open Set Automatic Modulation &#8230;The study investigates two prominent neural network architectures: the convolutional neural network (CNN) and attention mechanism based transformer.researchgate.net<\/a><a href=\"https:\/\/pmc.ncbi.nlm.nih.gov\/articles\/PMC11729949\/\" target=\"_blank\" rel=\"noreferrer noopener\">Modulation pattern recognition method of wireless communication &#8230;A signal modulation recognition method based on a two-way interactive temporal attention network algorithm has been developed.pmc.ncbi.nlm.nih.gov<\/a><a href=\"https:\/\/www.frontiersin.org\/journals\/communications-and-networks\/articles\/10.3389\/frcmn.2025.1453125\/full\" target=\"_blank\" rel=\"noreferrer noopener\">Modulation recognition method based on multimodal featuresThe method can effectively assign multi domain features, enabling the model to better recognize the signal.frontiersin.org<\/a><a href=\"https:\/\/ieeexplore.ieee.org\/iel8\/6287639\/10380310\/10568092.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">Automatic Modulation Open-Set Recognition Based on Random &#8230;The main application of deep learning-based automatic modulation recognition is to use neural network models to learn from one-dimensional.ieeexplore.ieee.org<\/a><a href=\"https:\/\/www.mdpi.com\/2079-9292\/13\/21\/4196\" target=\"_blank\" rel=\"noreferrer noopener\">SOAMC: A Semi-Supervised Open-Set Recognition Algorithm for &#8230;This paper proposes a semi-supervised open-set recognition approach, termed SOAMC (Semi-Supervised Open-Set Automatic Modulation Classification).mdpi.com<\/a><a href=\"https:\/\/www.sciencedirect.com\/science\/article\/abs\/pii\/S1874490725001946\" target=\"_blank\" rel=\"noreferrer noopener\">A comparative study of machine learning and deep learning for &#8230;Traditional modulation recognition (MR) techniques, employing likelihood-based (LB) and feature-based (FB) methods, struggle with accurate classification, &#8230;sciencedirect.com<\/a><\/p>\n\n\n\n<h3 class=\"wp-block-heading\">&#8220;MOS&#8221; &#8220;open set&#8221; detection OR OOD &#8220;RF&#8221; OR &#8220;signal classification&#8221; OR &#8220;modulation&#8221;10<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/iel8\/6488907\/11045559\/10943212.pdf\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/iel8\/6488907\/11045559\/10943212.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">A Deep-Learning-Based Open Set Automatic Modulation &#8230;<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/iel8\/6488907\/11045559\/10943212.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">3) We use group classifier and group constraint mecha- nism to enhance the open set modulation classification performance. Without the need to train AMC deep &#8230;<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ieeexplore.ieee.org\/iel8\/6488907\/11045559\/10943212.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">ieeexplore.ieee.org<\/a><a href=\"https:\/\/www.researchgate.net\/publication\/390287557_A_Deep_Learning_Based_Open_Set_Automatic_Modulation_Classification_Method_Using_Multiple_Domain_Representations_and_Group_Constraint\" target=\"_blank\" rel=\"noreferrer noopener\">A Deep Learning Based Open Set Automatic Modulation &#8230;Download Citation | A Deep Learning Based Open Set Automatic Modulation &#8230; MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Space.researchgate.net<\/a><a href=\"https:\/\/arxiv.org\/pdf\/2110.11334\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] Generalized Out-of-Distribution Detection: A Survey &#8211; arXivOOD Detection &amp; Open-Set Noisy Labels. Existing methods of learning &#8230; Li, \u201cMos: Towards scaling out- of-distribution detection for &#8230;arxiv.org<\/a><a href=\"https:\/\/www.sciencedirect.com\/science\/article\/abs\/pii\/S0263224124002677\" target=\"_blank\" rel=\"noreferrer noopener\">A novel high accuracy fast gas detection algorithm based on multi &#8230;Researchers have utilized MOS sensors to detect various gases &#8230; Qu et al. Open-set gas recognition: A case-study based on an &#8230;sciencedirect.com<\/a><a href=\"https:\/\/dl.acm.org\/doi\/10.1007\/s11263-024-02117-4\" target=\"_blank\" rel=\"noreferrer noopener\">Generalized Out-of-Distribution Detection: A SurveyMos: Towards scaling out-of-distribution detection for large semantic space. &#8230; open set recognition, arXiv preprint arXiv:2004.08067 \u00b7 Google Scholar. [128].dl.acm.org<\/a><a href=\"https:\/\/par.nsf.gov\/servlets\/purl\/10389819\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] A Self-Supervised Approach for Robust Out-of-Distribution DetectionMos: Towards scaling out-of- distribution detection for large semantic &#8230; Classification- reconstruction learning for open-set recognition. In Proceed &#8230;par.nsf.gov<\/a><a href=\"https:\/\/openreview.net\/pdf?id=DASh78rJ7g\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] PLUGIN ESTIMATORS FOR SELECTIVE CLASSIFICA &#8211; OpenReviewSurprisingly, this even holds in a special cases of OOD detection, such as open-set classification, wherein there is a strong relationship between Pin and Pout &#8230;openreview.net<\/a><a href=\"https:\/\/arxiv.org\/html\/2404.18279v1\" target=\"_blank\" rel=\"noreferrer noopener\">Out-of-distribution Detection in Medical Image Analysis: A surveyExtreme Value Theorem (EVT) is first applied to open set recognition by [71] . &#8230; Li, \u201cMos: Towards scaling out-of-distribution detection for &#8230;arxiv.org<\/a><a href=\"https:\/\/spj.science.org\/doi\/10.34133\/icomputing.0012\" target=\"_blank\" rel=\"noreferrer noopener\">Review on Algorithm Design in Electronic Noses: Challenges &#8230;For metal-oxide-semiconductor (MOS) sensors, the A\u2013D process is usually obtained through temperature modulation. Before discussing the temperature modulation &#8230;spj.science.org<\/a><a href=\"https:\/\/openaccess.thecvf.com\/content\/CVPR2025\/papers\/Wang_MetaShadow_Object-Centered_Shadow_Detection_Removal_and_Synthesis_CVPR_2025_paper.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">[PDF] MetaShadow: Object-Centered Shadow Detection, Removal, and &#8230;Task-Specific Datasets for Shadow Editing: We build a synthetic training set (MOS dataset) for shadow detection, &#8230; with cascaded modulation GAN and object-aware &#8230;openaccess.thecvf.com<\/a><\/p>\n\n\n\n<h3 class=\"wp-block-heading\">MOS method in open-set recognition or out-of-distribution detection10<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/355864561_MOS_Towards_Scaling_Out-of-distribution_Detection_for_Large_Semantic_Space\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/355864561_MOS_Towards_Scaling_Out-of-distribution_Detection_for_Large_Semantic_Space\" target=\"_blank\" rel=\"noreferrer noopener\">MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Space<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/355864561_MOS_Towards_Scaling_Out-of-distribution_Detection_for_Large_Semantic_Space\" target=\"_blank\" rel=\"noreferrer noopener\">In open-world recognition of safety-critical applications, providing reliable prediction for deep neural networks has become a critical requirement. Many methods have been proposed for reliable prediction related tasks such as confidence calibration, misclassification detection, and out-of-distribution detection. Recently, pre-training has been shown to be one of the most effective methods for improving reliable prediction, particularly for modern networks like ViT, which require a large amount of training data.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.researchgate.net\/publication\/355864561_MOS_Towards_Scaling_Out-of-distribution_Detection_for_Large_Semantic_Space\" target=\"_blank\" rel=\"noreferrer noopener\">researchgate.net<\/a><a href=\"https:\/\/link.springer.com\/chapter\/10.1007\/978-3-031-72658-3_27\" target=\"_blank\" rel=\"noreferrer noopener\">Operational Open-Set Recognition and PostMax Refinement | SpringerLinkIn: Conference on Computer Vision and Pattern Recognition (CVPR) (2018) &#8230; Huang, R., Geng, A., Li, Y.: On the importance of gradients for detecting distributional shifts in the wild. In: Advances in Neural Information Processing Systems (NeurIPS), vol. 34, pp. 677\u2013689 (2021) &#8230; Huang, R., Li, Y.: MOS: towards scaling out-of-distribution detection for large semantic space.link.springer.com<\/a><a href=\"https:\/\/github.com\/Jingkang50\/OpenOOD\/wiki\/OpenOOD-v1.5-methods-&amp;-benchmarks-overview\" target=\"_blank\" rel=\"noreferrer noopener\">OpenOOD v1.5 methods &amp; benchmarks overviewBenchmarking Generalized Out-of-Distribution Detection &#8211; OpenOOD v1.5 methods &amp; benchmarks overview \u00b7 Jingkang50\/OpenOOD Wikigithub.com<\/a><a href=\"https:\/\/github.com\/iCGY96\/awesome_OpenSetRecognition_list\" target=\"_blank\" rel=\"noreferrer noopener\">GitHub &#8211; iCGY96\/awesome_OpenSetRecognition_list: A curated list of papers &amp; resources linked to open set recognition, out-of-distribution, open set domain adaptation and open world recognitionExploring the Limits of Out-of-Distribution Detection. Stanislav Fort, Jie Ren, Balaji Lakshminarayanan. (ArXiv 2021). Can Subnetwork Structure be the Key to Out-of-Distribution Generalization?. Dinghuai Zhang, Kartik Ahuja, Yilun Xu, Yisen Wang, Aaron Courville. (ICML 2021). Out-of-Distribution Generalization in Kernel Regression. Abdulkadir Canatar, Blake Bordelon, Cengiz Pehlevan. (ArXiv 2021). MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Space.github.com<\/a><a href=\"https:\/\/arxiv.org\/abs\/2105.01879\" target=\"_blank\" rel=\"noreferrer noopener\">[2105.01879] MOS: Towards Scaling Out-of-distribution Detection for Large Semantic SpaceDetecting out-of-distribution (OOD) inputs is a central challenge for safely deploying machine learning models in the real world. Existing solutions are mainly driven by small datasets, with low resolution and very few class labels (e.g., CIFAR). As a result, OOD detection for large-scale image classification tasks remains largely unexplored.arxiv.org<\/a><a href=\"https:\/\/link.springer.com\/article\/10.1007\/s11263-024-02222-4\" target=\"_blank\" rel=\"noreferrer noopener\">Dissecting Out-of-Distribution Detection and Open-Set Recognition: A Critical Analysis of Methods and Benchmarks | International Journal of Computer VisionDetecting test-time distribution shift has emerged as a key capability for safely deployed machine learning models, with the question being tackled under various guises in recent years. In this paper, we aim to provide a consolidated view of the two largest sub-fields within the community: out-of-distribution (OOD) detection and open-set recognition (OSR).link.springer.com<\/a><a href=\"https:\/\/arxiv.org\/abs\/2105.01879v1\" target=\"_blank\" rel=\"noreferrer noopener\">[2105.01879v1] MOS: Towards Scaling Out-of-distribution Detection for Large Semantic SpaceDetecting out-of-distribution (OOD) inputs is a central challenge for safely deploying machine learning models in the real world. Existing solutions are mainly driven by small datasets, with low resolution and very few class labels (e.g., CIFAR). As a result, OOD detection for large-scale image classification tasks remains largely unexplored.arxiv.org<\/a><a href=\"https:\/\/openaccess.thecvf.com\/content\/CVPR2021\/papers\/Huang_MOS_Towards_Scaling_Out-of-Distribution_Detection_for_Large_Semantic_Space_CVPR_2021_paper.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">MOS: Towards Scaling Out-of-Distribution Detection for &#8230;openaccess.thecvf.com<\/a><a href=\"https:\/\/dl.acm.org\/doi\/10.1609\/aaai.v37i9.26230\" target=\"_blank\" rel=\"noreferrer noopener\">Towards in-distribution compatible out-of-distribution detection | Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-Fifth Conference on Innovative Applications of Artificial Intelligence and Thirteenth Symposium on Educational Advances in Artificial IntelligenceIn 2020 IEEE\/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2020, Seattle, WA, USA, June 13-19, 2020, 10948-10957. Computer Vision Foundation \/ IEEE. &#8230; Huang, R.; and Li, Y. 2021. MOS: Towards Scaling Out-of-Distribution Detection for Large Semantic Space.dl.acm.org<\/a><a href=\"https:\/\/openaccess.thecvf.com\/content\/CVPR2021\/html\/Huang_MOS_Towards_Scaling_Out-of-Distribution_Detection_for_Large_Semantic_Space_CVPR_2021_paper.html\" target=\"_blank\" rel=\"noreferrer noopener\">MOS &#8211; CVPR 2021 Open Access RepositoryMOS establishes state-of-the-art performance, reducing the average FPR95 by 14.33% while achieving 6x speedup in inference compared to the previous best method. &#8230; @InProceedings{Huang_2021_CVPR, author = {Huang, Rui and Li, Yixuan}, title = {MOS: Towards Scaling Out-of-Distribution Detection for Large Semantic Space}, booktitle = {Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, month = {June}, year = {2021}, pages = {8710-8719} }openaccess.thecvf.com<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Vote Tracing: Make RF Ensembles Auditable, Explainable, and Deployable TL;DR \u2014 Per-model \u201cvote traces\u201d turn your RF modulation ensembles into fully auditable systems: exact Shapley attribution in microseconds, gorgeous timelines, and an open-set detector that beats ODIN\/Energy without extra forwards. All the hooks live in classify_signal(); storage overhead is ~1\u20132 KB per signal. [Read the&hellip;&nbsp;<\/p>\n","protected":false},"author":2,"featured_media":3305,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"neve_meta_sidebar":"","neve_meta_container":"","neve_meta_enable_content_width":"","neve_meta_content_width":0,"neve_meta_title_alignment":"","neve_meta_author_avatar":"","neve_post_elements_order":"","neve_meta_disable_header":"","neve_meta_disable_footer":"","neve_meta_disable_title":"","footnotes":""},"categories":[10,11],"tags":[],"class_list":["post-4744","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-signal-science","category-signal_scythe"],"_links":{"self":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4744","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=4744"}],"version-history":[{"count":0,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4744\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=\/wp\/v2\/media\/3305"}],"wp:attachment":[{"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4744"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4744"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/neurosphere-2.tail52f848.ts.net\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4744"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}