arXiv:2609.38956v1 Announce Type: new Abstract: Routing signals of modern vision transformers -- expert gates, attention-residual weights and halting scores -- often improve probes that predict wheth...
The paper presents a validated protocol for adapting drone‑based crowd‑counting models to the extreme conditions expected at the 2034 FIFA World Cup in Saudi Arabia. Using 525 controlled runs and a full‑resolution corpus, the authors demonstrate that label‑free adaptation can recover 31‑49% of shift‑induced error across multiple corruptions and severities, achieving a 41.8 MAE improvement over a frozen source model. They also introduce a severity law, a stability budget, and a flux‑based risk module that detects real congestion episodes, culminating in a six‑point deployment protocol for safe aerial crowd monitoring.
arXiv:2605.00915v2 Announce Type: replace Abstract: Probing frozen vision transformers typically uses permutation-invariant aggregation (GAP or $\texttt{[CLS]}$), treating patch tokens as an unstruct...
VeriPhy is an auditable physical‑verification system that evaluates generated video by compiling prompts into typed physical obligations and a statically validated execution plan before any frames are observed. During execution, it gates calls to frozen low‑level experts (e.g., segmentation, tracking, counting, depth, OCR, audio‑event detection) and returns provenance‑carrying evidence records, which are mapped to a three‑valued state (supported, contradicted, unknown) with full traceability. On a 1,500‑clip corpus of human‑annotated flaw records, VeriPhy accounts for 228 failures out of 304, outperforming a published question‑decomposition evaluator that accounts for 164, while also providing auditable evidence for each verdict.
VeriPhy is an auditable physical‑verification system that transforms a text prompt into typed physical obligations and a statically validated execution plan before any video frames are generated. During execution, it gates calls to frozen low‑level experts (segmentation, tracking, counting, depth, OCR, audio‑event detection, etc.) and records provenance‑carrying evidence for each action. The system maps these records to a three‑valued state—supported, contradicted, or unknown—providing traceable verdicts that can be used to refine generation models.
arXiv:2608. 02464v1 Announce Type: cross Abstract: LLM agents fail mid-episode -- they loop, cascade tool errors, drift off goal, fabricate results, or silently absorb corrupted content -- and the standard remedy, judging every step with a second LLM, costs more than the agent itself.
arXiv:2608. 02302v1 Announce Type: cross Abstract: Long-horizon coding-agent trajectories are poorly matched to the credit units available to train on: a single action has no stable value, an episode label merges productive exploration with abandoned directions, and a fixed window cuts where the logging mechanics fall.
arXiv:2608. 10145v1 Announce Type: new Abstract: LeWorldModel trains a latent world model with a prediction loss and a single anti-collapse regulariser, and reports approximately 87% of goals reached on TwoRoom, its simplest diagnostic environment.
arXiv:2608. 06361v1 Announce Type: new Abstract: Real-world video benchmarks provide broad coverage, but their fixed clips entangle event count, rate, duration, and visual complexity, making failure modes hard to isolate.
TrajMind is a framework for diagnosing collective anomalies in urban trajectory data. It separates continuous screening from on-demand diagnosis, using a fast text-only path for alerts and a slow vision‑language path that chains role‑specialized LoRA adapters for detailed, evidence‑backed what‑who‑where‑when records. Experiments show the slow path outperforms baselines by over 15 percentage points in typing and 13 in localization, while the fast path cuts latency by 41% and retains high accuracy.
arXiv:2609.37603v1 Announce Type: cross Abstract: A common safeguard for a data pipeline is redundant computation: derive each published number by two routes built on different technology and refuse...
DeFA is a dependency-guided framework that attributes failures in large language model agents by constructing an event dependency graph and a failure propagation graph from protocol relations and semantic dependencies. It identifies violating events, traces their sources and effects, and determines the decisive error, responsible agent, and error category. The method supports long trajectories through segmentation and has shown superior accuracy on text, image, and video tasks, while its diagnostic feedback can improve agent performance on subsequent tasks.