{"slug": "causal-neural-set-filtering-for-online-multi-target-tracking", "title": "Causal neural set filtering for online multi-target tracking", "summary": "Researchers proposed Causal Neural Set Filtering (CNSF), a neural set filter that encodes only current measurements while carrying past evidence in a structured recursive track state, reducing mean GOSPA and T-GOSPA relative to Track-MT3 by 19.3% and 30.4% on a held-out three-regime simulated test set. CNSF combines exclusive Sinkhorn association, association-conditioned Kalman-shaped updates with moment matching, and recurrent Bernoulli lifecycle modeling with measurement-driven birth, achieving 55.9% fewer parameters and a 3.76x speedup in single-thread CPU inference. The work targets redundant computation in Transformer-based multi-target tracking, where MT3/Track-MT3-style trackers repeatedly re-encode measurement windows.", "body_md": "arXiv:2609.16054v1 Announce Type: new \nAbstract: Transformer-based multi-target tracking (MTT) jointly learns data association and state estimation, but MT3/Track-MT3-style trackers repeatedly re-encode measurement windows, incurring redundant computation. We propose Causal Neural Set Filtering (CNSF)\\footnote{\\href{https://github.com/daihuangyu/CNSF}{Code: https://github.com/daihuangyu/CNSF}}, a neural set filter that encodes only current measurements while carrying past evidence in a structured recursive track state. CNSF combines exclusive Sinkhorn association, association-conditioned Kalman-shaped updates with moment matching, and recurrent Bernoulli lifecycle modeling with measurement-driven birth. These mechanisms impose soft one-to-one constraints, propagate association-induced state uncertainty, and support existence estimation under missed detections and birth--death transitions. On a held-out three-regime simulated test set, CNSF reduces mean GOSPA and T-GOSPA relative to Track-MT3 by 19.3\\% and 30.4\\%, with 55.9\\% fewer parameters and a $3.76\\times$ speedup in single-thread CPU inference.", "url": "https://wpnews.pro/news/causal-neural-set-filtering-for-online-multi-target-tracking", "canonical_source": "https://arxiv.org/abs/2609.16054", "published_at": "2026-09-16 04:00:00+00:00", "updated_at": "2026-09-16 04:06:43.078465+00:00", "lang": "en", "topics": ["machine-learning", "ai-research", "neural-networks"], "entities": ["Causal Neural Set Filtering", "Track-MT3", "MT3", "Sinkhorn association", "GitHub"], "alternates": {"html": "https://wpnews.pro/news/causal-neural-set-filtering-for-online-multi-target-tracking", "markdown": "https://wpnews.pro/news/causal-neural-set-filtering-for-online-multi-target-tracking.md", "text": "https://wpnews.pro/news/causal-neural-set-filtering-for-online-multi-target-tracking.txt", "jsonld": "https://wpnews.pro/news/causal-neural-set-filtering-for-online-multi-target-tracking.jsonld"}}