Papers

Reconstructing visual stimuli from electroencephalography (EEG) is difficult because scalp measurements have high temporal but limited spatial resolution, and paired EEG-image datasets remain small relative to modern generative-model training corpora. We present a reproducible single-subject baseline on THINGS-EEG2 that first tests the more defensible question of whether EEG can retrieve the viewed stimulus in a visual embedding space. A compact temporal-spatial convolutional encoder maps repetition-averaged EEG (63 by 250) to provided 512-dimensional ViT-B/32 image features. Model selection uses a concept-disjoint validation split, and final evaluation uses the official 200-image, 200-concept test gallery. Across three training seeds, the model obtains 12.83 +/- 0.58%, 39.17 +/- 1.76%, and 58.00 +/- 1.73% image recall at 1, 5, and 10 (mean +/- sample standard deviation), compared with analytical chance levels of 0.5%, 2.5%, and 5.0%. A session-balanced ablation shows that averaging more test repetitions generally improves ranking. Applying the Subject 01 model to the other nine subjects without adaptation causes a sharp performance drop, exposing subject specificity. We further report exploratory stress tests of direct conditional generators trained without external visual weights: single-subject and ten-subject variants produce noise-dominated outputs, with early validation improvements reversing after one to four epochs. Finally, we distinguish direct reconstruction from semantic rendering with a pretrained diffusion prior. The results support above-chance coarse semantic decoding under a closed-set, repetition-averaged protocol, but do not support faithful recovery of stimulus pixels.

Foundation models have recently demonstrated strong capabilities across a wide range of medical imaging tasks. However, their performance in structured clinical interpretation settings remains insufficiently explored. In lung cancer screening, interpretative variability persists despite standardized frameworks such as Lung-RADS. In this study, we evaluate MedGemma, a medical general-purpose foundation model derived from Gemini and its fine-tuned version adapted for lung cancer detection and diagnosis, compared against radiologists performing Lung-RADS v2022 assessment on the NLST dataset. Twelve radiologists independently evaluated each case in a multi-reader design, enabling quantification of inter-reader variability. Radiologists achieved a mean AUC of 0.90, with substantial variability across readers (range: 0.80-0.94). The native foundation model achieved an AUC of 0.70, failing to reach clinically relevant performance. In contrast, fine-tuning significantly improved performance to an AUC of 0.83, placing the model within the lower range of individual radiologists performance. These findings highlight a trade-off between peak accuracy and prediction consistency. Unlike radiologists, under fixed conditions, the model produces deterministic outputs, removing inter-run variability under identical inputs, in contrast to inter-reader variability observed among radiologists. This supports the role of fine-tuned foundation models potential complementary tools for clinical decision support, particularly in settings with limited expertise. However, evaluation is performed on a case-enriched cohort from NLST and does not account for real-world prevalence or external validation, limiting direct clinical generalization.

Computational text-based ideal point estimation (CT-IPE) methods are usually compared as named algorithms, yet applying them involves numerous researcher choices that configure how political text is turned into position estimates. This paper argues that CT-IPE methods are better understood as configurable measurement pipelines than as fixed estimators. Building on a large-scale comparative experiment spanning 17 CT-IPE algorithms, 5,537 experimental runs, and approximately 4.25 million left-right position estimates, I describe the shared infrastructure that makes these heterogeneous methods jointly executable and quantify how sensitive their estimates are to alternative hyperparameter choices. Variance-partitioning and SHAP-based sensitivity analyses show that, for most algorithms, hyperparameter profiles explain little residual variance through a shared shift: 13 of the 17 algorithms exhibit ICC values below .10. Where this profile-level sensitivity is present, it is concentrated in a small number of consequential researcher choices, most notably the selection of the underlying language or embedding model, the seed keyword lists that anchor the construct, and the number of topics.

This paper studies decentralized learning of socially optimal equilibria in finite normal-form games over dynamic communication networks. Each agent observes only its own realized payoffs, does not know the game a priori, and can communicate only with time-varying neighbors using low-bandwidth messages. We propose networked decentralized optimal equilibrium learning dynamics in which agents generate randomized semantic content/discontent signals from local payoff comparisons and exchange time-stamped time-stacked tables rather than raw actions, payoff information or local estimates/parameters. The method combines table fusion with temporal majority reconstruction to mitigate dynamic communication while preserving fully decentralized operation. We establish finite-time logarithmic regret guarantees, with an in-phase exploration perturbation, for optimal equilibrium selection under utilitarian and proportional-fair social welfare objectives. Simulation results further show that the proposed approach can effectively select socially desirable equilibria over dynamic communication networks.

A small number of unusually high-gain parameters can exert disproportionate effects in transformer language models, but whether analogous structures recur in genomic foundation models and whether structural geometry determines functional importance remains unknown. We analyzed high-gain rows in gated feed-forward networks across text and genomic foundation models, including a frozen 22-model causal census. Computing an associated bilinear weight operator exactly, without a diagonal approximation, we tested whether structural extremeness is a transferable mechanism. Activation-derived candidates were functionally enriched relative to random and top-norm same-layer controls, yet neither spectral concentration nor operator magnitude predicted causal effect size, and these associations vanished within the endpoint-homogeneous text-decoder subset. A within-layer sweep of 36 rows in one genomic and one text decoder resolved this into two regimes: below the detector's acceptance threshold the ratio carried no positive information about causal damage, whereas above it the ratio ordered rows strongly but did not grade severity as a dose-response. The same sweep revealed a second individually catastrophic row invisible to a one-candidate-per-model census, and non-additive damage among co-located critical rows. Case studies showed divergent causal organizations: a robust super-additive pair interaction in DNABERT-2, and in GENERator a sharply position-localized dependence in which preserving or restoring the row's beginning-of-sequence contribution rescued essentially all native-loss damage. High-gain gated-FFN rows are therefore a recurrent architectural phenotype whose structural prominence acts as an enrichment signal, not a calibrated measure of functional criticality or a specification of causal organization. Enrichment is general, but the mechanism is model-specific.

Property-based testing (PBT) with QuickChick relies on extracting Rocq programs to OCaml. However, this extraction mechanism, while crucial for QuickChick to function, has significant performance implications. In this work, we describe how we optimized QuickChick, exploiting various opportunities offered by program extraction to significantly improve each test's extraction, compilation, and running time. We also evaluate these improvements using the ETNA benchmarking platform for PBT to assess how individual improvements impacted the overall test run time.

The binary length function $\ell_2(r,R)$ is the smallest length of a binary linear code with codimension (redundancy) $r$ and covering radius $R$. Let $s_2(N,\rho)$ be the smallest size of a $\rho$-saturating set in the binary projective space $\mathrm{PG}(N,2)$. It is known that the binary length function satisfies $\ell_2(r,R)=s_2(r-1,R-1)$. We obtain the following new upper bounds on $\ell_2(r,R)$, improving the previously known upper bounds by $\Delta(r,R)$: $R=2,r=2t\ge28$, and $r=10,18,20$, $\ell_2(r,2)=s_2(r-1,1)\le51\cdot2^{r/2-5}-1,\Delta(r,2)=2^{r/2-5}$; $R=3,r=3t-1\ge41$ and $r=26$, $\ell_2(r,3)=s_2(r-1,2)\le818\cdot2^{(r-26)/3}-1$, $\Delta(41,3)=63,\Delta(r,3)=2^{(r-26)/3}$ if $r\ge44$. To obtain these bounds, we construct new infinite code families, using distinct versions of the $q^m$-concatenating constructions of covering codes, including new versions proposed in this paper. We also obtain new useful partitions of column sets of parity check matrices of some codes. The new code families provide the asymptotic covering-density bounds $\overline{\mu}(2)\le2601/2048<1.27002$, $\overline{\mu}(3)\le1.35935$, improving the previously known bounds and giving rise to the new upper bounds $f(2)\le2601/2048<1.27002$, $f(3)\le1.35935$ on the constants $f(2),f(3)$ of the Green's Open Problem 40.

Physiological affect sensing in naturalistic group interaction is often limited by sparse labels rather than sensor data: wearable devices produce many time windows, while self-reports are collected only a few times per session. Using GroupAffect-4, a four-person collaborative dataset with wearable physiology, eye tracking, Big Five personality, and post-task VAD labels, we study pseudo-label augmentation for affect sensing under sparse supervision. We compare no augmentation, Gaussian Process pseudo-labelling, personality-aware trust weighting, and joint personality-plus-confidence weighting within a shared target-construction pipeline. Results show that pseudo-label augmentation improves over the labelled-only baseline in the known-team setting. However, the narrow range of Big Five cosine similarities (0.91-0.99) makes fine-grained personality weighting ineffective; personality similarity functions mainly as a same-team filter rather than a calibrated trust signal. With smoothing, augmented SVM variants are effectively tied on Valence and Arousal, while the joint personality-plus-confidence variant gives the highest Dominance score. Cross-subject LOSO transfer remains encouraging, especially for Arousal, whereas strict session-isolated LOGO removes the augmentation benefit. Given only 10 groups, LOGO should be interpreted as a conservative lower bound on unseen-group transfer. Overall, the results suggest that pseudo-label augmentation can make better use of sparsely labelled collaborative affect data, while personality information is most useful as a within-team selection mechanism.

Large Language Models (LLMs) excel at programming tasks but frequently fail at deterministic, fine-grained reasoning in natural language, relying heavily on semantic approximations rather than robust symbolic execution. To bridge this gap, we propose MIMIC, a framework that leverages executable code as a rigorous medium for reasoning data synthesis. MIMIC fundamentally transforms algorithms into verifiable reasoning trajectories through narrative fusion, code-guided test synthesis, and dynamic code instrumentation. Crucially, these explicit intermediate execution states naturally form a Code-Instrumented Reward (CIR), providing dense, high-fidelity process supervision for reinforcement learning without external reward models. Extensive evaluations reveal that models trained via SFT and GRPO on our synthesized dataset achieve substantial, consistent gains. Our method significantly elevates accuracy across general reasoning, complex mathematical benchmarks, and fine-grained deterministic tasks, demonstrating that the procedural rigor of executable code can effectively unlock and enhance the generalized reasoning capabilities of LLMs. Our code and data are available at https://github.com/zjy1298/MIMIC.

Flexible robotic production requires joint decisions on process progression, material routing, resource assignment, temporary cooperation, and simultaneous execution, since each decision can affect the feasibility of the others. The challenge is greater under decentralized control, where each robot acts from bounded local information while system progress depends on collective decisions, shared resources, material state, and workspace compatibility. These properties closely match cooperative multi-agent decision making under partial observability and resource contention. This paper introduces AssemblyGrid v1, a reproducible benchmark for repeated multi-robot production that combines explicit process progression, decentralized observations, material transfer, temporary multi-robot coalitions, productive concurrency, and geometry-dependent feasibility within one task-level formulation. The benchmark includes Flow, Coalition, and Concurrency workload families, each with three scenario levels. Task success and evaluation measures are defined independently of learning reward and solution method, allowing learning-based and non-learning methods to address the same production problem. AssemblyGrid v1 is evaluated through executable conformance checks, mechanism studies, and algorithmic experiments using a privileged centralized reference, structured decentralized controllers, and MARL methods including IPPO, MAPPO, and QMIX. Results demonstrate productive execution under centralized and decentralized control. The MARL experiments further show that decentralized policies can learn effective production behavior from local observations and actions, supporting AssemblyGrid as a controlled benchmark for studying cooperative decision making in flexible robotic production.

Robots operating in open environments act under partial observability, physical constraints, and dynamic task contexts. Beyond mapping observations and language instructions to actions, they must anticipate how candidate actions may affect future states and task-relevant outcomes. Recent advances in world models, video generation, and Vision-Language-Action (VLA) policies have motivated the development of World-Action Models (WAMs), which couple future world prediction with executable action generation. This survey provides a robotics-oriented review of WAMs. We clarify their scope relative to conventional world models, model-based reinforcement learning, action-conditioned video generation, and reactive VLA policies, and organize existing methods through a unified taxonomy covering representations, transition modeling, action interfaces, architectures, training pipelines, data modalities, and scaling strategies. We further review applications of WAMs in manipulation, navigation, and autonomous driving, and we summarize the datasets, benchmarks, metrics, and protocols used to evaluate WAM systems. Finally, we discuss key challenges in action alignment, world-action factorization, spatial and multi-view consistency, long-horizon memory, neural simulation for closed-loop policy learning, and efficient inference. Taken together, this survey aims to provide a concise technical foundation for integrating predictive world modeling with action generation, toward more reliable embodied robot intelligence. Project page: https://rcl-robotics.github.io/Awesome-World-Action-Models.

Retrieval-Augmented Generation (RAG) serves as the primary memory architecture for long-horizon autonomous agents. However, treating shared memory as an append-only stream introduces Semantic Shadowing, a critical failure mode where conflicting historical observations accumulate and statistically dominate valid recent updates. In dynamic environments, this results in severe state divergence as agents retrieve and act upon obsolete facts. This paper formalizes the mechanics of State Mutability to prove that standard dense retrieval suffers from Asymptotic Recall Decay. Furthermore, we formally demonstrate a Majority Vote Trap, revealing that increasing the retrieval context window paradoxically degrades generation accuracy by diluting the attention mechanism under conditions of semantic equivalence. To resolve this, we introduce GC-Mem (Garbage Collection for Memory), a strict inference-time consistency protocol. Unlike heuristic time-decay mechanisms—which indiscriminately destroy valid long-term memory—GC-Mem relies purely on a temporal dominance operator ($\Phi_{\mathcal{T}}$) paired with contradiction detection to surgically excise shadowed context. Evaluated across a rigorous, behaviorally inferred benchmark of 137,760 memory chunks and continuous accumulation sweeps, standard RAG and timestamp re-ranking baselines experience severe degradation. In contrast, GC-Mem empirically recovers $>90\%$ conflict resolution accuracy. We establish strict precision and recall deployment thresholds, ensuring state convergence where standard mutable RAG fundamentally fails.

Industrial world models must move between machines whose dynamics, sensing interfaces and command conventions differ. This study asks whether a command-conditioned latent world model, trained to predict future representations of the process rather than to reconstruct future samples, keeps its value on a machine it has never seen: a source CNC machine exposes 17 sensor channels, the target sharing 10 of those. All model selection uses source data only, and the locked configuration is evaluated on the target once. Two findings follow. First, latent-predictive pretraining brings no in-domain forecasting gain over matched training from scratch, so source accuracy alone cannot show what such a representation is worth. Second, the transferred model beats persistence on the unseen machine (with $R^2\approx0.01$ against the target mean) but trails official forecasters that normalize each input window by its own statistics; a post-lock ablation, declared before it ran, shows that this input normalization alone closes the gap, and closing it costs predictive calibration. Cross-machine transfer under partial sensor overlap is therefore a distinct evaluation axis for command-conditioned world models.

Generative recommendation reformulates item prediction as semantic identifier generation, yet episodic content introduces a fundamentally different setting where the target is determined by narrative evolution rather than user preference. This task requires models to understand multimodal storyline progression while addressing the efficiency challenges caused by redundant visual contexts and costly explicit reasoning generation. We propose NarraLite, an efficient multimodal generative recommendation framework that jointly compresses perception and reasoning. Specifically, Progressive Spectral Compression selectively distills long visual contexts into compact narrative-relevant evidence, preserving transition-critical information while reducing redundant visual computation. Latent Narrative Reasoning introduces context-routed latent reasoning tokens and aligns their contextualized representations with future continuation semantics, enabling implicit narrative inference without autoregressively decoding textual rationales. We further establish a user-agnostic multimodal benchmark for short-form drama continuation across UGC, PGC, and OOD settings. Extensive experiments demonstrate that NarraLite consistently improves continuation accuracy, narrative coherence, and robustness over existing approaches, while achieving a favorable accuracy–efficiency trade-off.

Synthetic tabular data can match real data distributions while still violating the semantic constraints that govern valid tabular rows. This reveals a key limitation of existing tabular generators: they mainly optimize distributional fidelity, but do not explicitly model weak semantic priors encoded in tabular schema and textual descriptions. In this paper, we propose \ours, a semantics-consistent tabular diffusion framework for high-fidelity synthetic data generation under weakly specified semantic priors. \ours\ first constructs two types of priors, namely intra-column semantics and inter-column symbolic rules, with LLM-assisted extraction from metadata and validation on the real training split. These priors are then used as generation conditions rather than post-hoc filters. Specifically, \ours\ maps heterogeneous column values, column identities, and semantic priors into a unified semantic space, and performs column-wise forward corruption and prior-conditioned reverse denoising to preserve both marginal distributions and rule-consistent cross-column dependencies. Extensive experiments on six real-world tabular benchmarks show that \ours\ consistently improves distributional fidelity, semantic consistency, and downstream task utility over representative VAE-, GAN-, LLM-, and diffusion-based baselines. Additional analyses further demonstrate the robustness of \ours\ when semantic priors are partially unavailable.

In order to solve the high-dimensional ($d \geq 30$) expensive black-box problems within budget, an efficient global optimization (EGO) algorithm with a dynamic aggregation strategy is proposed, labeled as DA-EGO. Specifically, the DA-EGO decomposes the original high-dimensional design space into a set of low-dimensional subspaces for efficient surrogate-based optimization search, and the optimal solutions of subspaces are combined as an elite point for the global search. Most importantly, the subspaces are not fixed. Instead, the subspace variables are updated in each iteration, according to the variable interaction analyses in the sub- and full-spaces. The perturbation method and the analysis of variance are used to detect variable interactions. To further accelerate the optimization progress, the searching ranges of subspaces are also adaptively adjusted according to the analyses of subspace optimization results of the previous iteration. Tests on 21 benchmark instances, comprising seven functions at 30, 60, and 90 dimensions, show that DA-EGO is effective on separable and partially separable problems under a budget of 1500 function evaluations. Its advantage is case-dependent: on the non-separable shifted Rosenbrock function, GSGA performs better at 60 and 90 dimensions, while the 30-dimensional results are statistically comparable to IKAEA and GSGA. Moreover, the advantage of DA-EGO is also seen in the aerodynamic optimization of a transonic rotor blade with 28 variables as well as the compressor stage optimization with 60 variables. With the above, the effectiveness of the proposed DA-EGO has been well demonstrated.

To enable flexible and rapid aerodynamic performance evaluation in turbomachinery design, this paper proposes a panoramic performance prediction framework. Unlike most previous prediction models that directly predict the objective functions of interest, our approach first predicts the basic parameters of the Navier-Stokes equations, such as temperature, pressure, and density. Utilizing these basic physical quantities, it subsequently predicts key performance parameters of the turbine stage meridian plane. By adopting this methodology, our proposed panoramic performance prediction framework functions similarly to a CFD simulator, capable of predicting various objective of interest to the designers. To enhance prediction accuracy, a transformer-enhanced neural operator (TNO) is introduced within this framework. Using the Rotor 37 blades as a reference, the proposed TNO is trained to predict the performance of a transonic compressor blade in the meridian plane. The TNO can accurately predict total quantities such as isentropic efficiency, mass flow, and distributions of total pressure ratio. Remarkably, the prediction error of TNO is observed to be smaller than that of state-of-the-art deep learning operators such as the FNO and DeepONet. Furthermore, the TNO is applied to downstream tasks, including sensitivity analysis and optimization of various objective functions. The results confirm that the TNO can operate almost like a CFD simulator, while reducing the computational cost of downstream tasks by four orders of magnitude. The effectiveness and reliability of the proposed TNO for solving different kinds of downstream tasks have been well demonstrated.

Human activity recognition (HAR) is usually framed as gradient-based training of neural networks. Agentic Heuristic Learning (AHL) Studio explores a complementary view inspired by human cognitive learning: people learn activities by remembering examples, forming rules, and repairing mistakes, not by backpropagating. This proposed tool implements AHL for HAR: a learning-time agent reasons over sensor protocols, proposes executable heuristic policies, records repair traces, and exports an LLM-free policy for edge deployment. We focus on the HAR benchmark family and provide an end-to-end workflow from dataset observation to edge-oriented export. On eleven HAR datasets evaluated so far, AHL policies reach strong executable-policy performance while remaining inspectable, editable, and replayable \footnote{https://github.com/zhaxidele/ahl-ts-studio}.

The array programming paradigm applies operations to entire arrays rather than individual elements, eliminating explicit loops and abstracting many low-level details of computation. Two influential approaches have shaped array programming: NumPy's broadcasting and Kenneth Iverson's array-oriented notation, with the latter first introduced in APL and later developed in the J language. Although NumPy's array programming model was influenced by Iverson's work, the two systems formalize array operations differently: NumPy primarily through array shapes and broadcasting, and J through verb rank. This paper establishes a formal correspondence between NumPy broadcasting and J rank and presents an implementation of NumPy broadcasting in J. The results provide a formal connection between two influential approaches to array-oriented computation and show how they address the common problem of applying operations to arrays with differing shapes without explicit loops.

We study signed, weighted affine $p$-adic residual objectives as native encodings of finite-domain constraints. For primes that separate the finite alphabet, sufficiently weighted positive unary rows pin each coefficient to its allowed set, while negative rows reward unequal endpoints or clause satisfaction. A coordinatewise domination theorem places every global minimiser in the finite domain; there the loss is, up to an additive constant, the all-different conflict count or the negative number of satisfied CNF clauses. Standard Sudoku provides an $81$-coefficient case study without a one-hot lift. A client-side implementation exposes the generated dataframes, arithmetic, diagnostics, and searches.

The studying of opinion dynamics and its propagation within social networks is crucial for addressing a wide range of challenges, including political polarization, public health, and marketing strategies. In this work, we study the problem of opinion dynamics by proposing a framework based on Friedkin-Johnsen (FJ) to identifies influential users and study their impact on dynamics opinions of community. The FJ model assume each individual have two opinions: initial and expressed. Through a series of initial opinion manipulation experiments, the proposed framework assesses the impact of influential versus random users on the overall community opinion. The proposed framework is validated using a tweet dataset representing the U.S. presidential election. The results shows that influencers with highest influencing score, significantly shift the overall community opinion. Moreover, the results shows that the impact of influencers not limited to direct neighbors , but beyond it, to their neighbors of neighbors . This study demonstrates how digital influencers on social media can shape public opinion regarding a subject or cause.

Span-level masked language modeling (MLM) has shown to be advantageous to pre-trained language models over the original single-token MLM, as entities/phrases and their dependencies are critical to language understanding. Previous works only consider span length with some discrete distributions, while the dependencies among spans are ignored, i.e., assuming that the positions of masked spans are uniformly distributed. In this paper, we present POSPAN, a general framework to allow diverse position-constrained span masking strategies via the combination of span length distribution and position constraint distribution, which unifies all existing span-level masking methods. To verify the effectiveness of POSPAN in pre-training, we evaluate it on the datasets from several NLU benchmarks. Experimental results indicate that the position constraint is capable of enhancing span-level masking broadly, and our best POSPAN setting consistently outperforms its span-length-only counterparts and vanilla MLM. We also conduct theoretical analysis for the position constraint in masked language models to shed light on the reason why POSPAN works well, demonstrating the rationality and necessity of POSPAN.

Users often need ask questions and seek answers online. The Question - Answering (QA) forums such as Stack Overflow cannot always respond to the questions timely and properly. In this paper, we propose HintMiner, a novel automatic question hints mining tool for users to help them find answers. HintMiner leverages the machine comprehension and sequence generation techniques to automatically generate hints for users' questions. It firstly retrieve many web Q&A posts and then extract some hints from the posts using MiningNet that is built via a language model. Using the huge amount of online Q&A posts, we design a self-supervised objective to train the MiningNet that is a neural encoder-decoder model based on the transformer and copying mechanisms. We have evaluated HintMiner on 60,000 Stack Overflow questions. The experiment results show that the proposed approach is effective. For example, HintMiner achieves an average BLEU score of 36.17% and an average ROUGE-2 score of 36.29%. Our tool and experimental data are publicly available.

Multimodal instruction following (MMIF) is crucial for building generalist agents. However, current training paradigms rely heavily on Supervised Fine-Tuning (SFT), which often leads to surface-level pattern matching and degrades general capabilities. While Reinforcement Learning with Verifiable Rewards (RLVR) offers a promising alternative, its scalability in MMIF is severely bottlenecked by the scarcity of high-quality, RL-ready multimodal data. To bridge this gap, we present MIFS (Multimodal Instruction Following Synthesis), a systematic pipeline designed to generate RL-ready multimodal data. Specifically, MIFS introduces a generative constraint protocol to synthesize diverse raw samples, followed by a learnability-aware distillation mechanism that filters data based on RL training dynamics to ensure stable policy optimization. Furthermore, a code-based verifier provides high-precision reward signals for policy learning. The resulting dataset comprises 90k samples across 8 constraint categories and 14 task domains. Empirical evaluations demonstrate that MIFS-trained MLLMs achieve an average improvement of 8.13% on four MMIF benchmarks and a 3$\times$ faster training convergence compared to using raw data. Crucially, our approach mitigates the generalization trade-offs typical of SFT, preserving core visual capabilities while significantly boosting instruction-following precision.

Red-light violations and harsh braking at signalized intersections are major contributors to traffic accidents. This paper analyzes and predicts human driver decision-making and longitudinal trajectory behavior during traffic light signal transitions. We collected a diverse real-world dataset comprising 449 approach runs under varying speed and distance conditions. Vehicle motion was recorded using RTK-corrected GNSS with centimeter-level accuracy, and driver heart rate and multi-level comfort ratings were monitored. Spatial and temporal calibration ensured precise alignment between vehicle state and signal timing. Statistical analysis identifies required deceleration as the dominant single predictor of the stop-go decision, and heteroscedastic Gaussian modeling of peak deceleration reveals five empirical comfort ranges derived from human stopping behavior. Based on this insight, we propose a two-stage modeling framework. Stage 1 predicts the binary maneuver decision, and Stage 2 generates the longitudinal acceleration trajectory using a decision-conditioned autoregressive Transformer with physics constraints, including target-state conditioning and jerk limits. The proposed architecture outperforms baseline methods and achieves 0.49m/s^2 acceleration MAE and 0.62m distance MAE. It also estimates the future stopping-comfort level of the human driver from a single yellow-onset snapshot. Qualitative results demonstrate realistic human-like braking behavior. The dataset and source code are publicly available.

Unified multimodal large language models (MLLMs) and multi-agent systems have advanced visual generation. However, three limitations remain. (1) Existing methods often distill task-specific experience with limited generalizability. (2) Reflection is often deferred until task completion. (3) Knowledge is often acquired only in response to downstream task demands. To address these limitations, we introduce OmniHarness, a framework for generalizable visual generation via symbolic policy learning. OmniHarness abstracts verified executions into symbolic policies for visual generation task families, capturing shared procedures and applicability conditions while removing instance-specific inputs. The harness instantiates, adapts, and composes these policies for new tasks. Intermediate verification guides refinement and failure recovery during execution. Through self-directed inquiry, OmniHarness autonomously generates and executes practice tasks near its capability limits before downstream objectives are specified. Execution feedback continually refines the policies while model parameters remain fixed. Experiments across six benchmarks, three MLLM backbones, and three visual agent frameworks demonstrate strong performance and continual capability expansion. On ComfyBench's Creative tasks, OmniHarness achieves a 95.0% resolve rate, exceeding the strongest baseline by 27.5 percentage points. Frozen policy snapshots improve existing visual agent systems through plug-and-play reuse.

Humans carry behaviour knowledge of how to act in familiar situations into every new task rather than relearning it from scratch. There is no reason a Reinforcement Learning (RL) agent shouldn't do the same: known behaviour patterns need not be learned, only applied. Neuro-symbolic RL bridges prior knowledge and RL by injecting symbolic knowledge alongside a learned policy. The point at which this knowledge is integrated is critical: a poor choice can produce, for instance, hallucinated preconditions, which surface as safety and reliability problems in agents acting in changing environments. We formalise this behavioural knowledge as a precondition Bayesian network (BN) over the agent's structural actions - the actions whose legality depends on preconditions, such as picking up a key, grasping a block, toggling a door, or dropping an object. The BN restricts when these actions may fire, and we inject it into the RL loop at three placements: (1) a symbolic verifier, consulted only at inference, that fires a structural action once its preconditions hold; (2) a symbolic enforcer, active during both training and inference, that governs structural-action use throughout learning; and (3) a symbolic learner, which folds the knowledge into the network and learns the restriction and use of structural actions itself. To test the three variants we run experiments on two benchmarks with opposite regimes: one built on long, ordered planning chains, the other on continuous manipulation. We compare against strong baselines on solution quality, sample efficiency, and traceability. The payoff is substantial. On MiniGrid, all three placements improve the solution quality over the PPO+RND baseline, the symbolic enforcer leading at $98.2\%$ against the baseline's $88.8\%$. On Fetch, $\dots$

Test-time compute has emerged as a major approach to improving the capabilities of Large Language Models (LLMs). However, existing test-time reasoning paradigms rely heavily on externally imposed control, either through fixed reasoning programs or through costly expansion in constrained search spaces, limiting both generalization and efficiency. We propose State of Thought (SoT), a new reasoning paradigm that enables endogenous reasoning in LLMs, with the model's internal reasoning state governing how reasoning unfolds. Concretely, SoT extracts a compact dynamics-geometric state from the model's internal information transfer and uses a 582-parameter controller on frozen backbones to selectively activate historical reasoning support useful under the current reasoning state, framing reasoning as a state-conditioned process over evidence rather than an externally prescribed token chain. Across quantitative (1.34x), general (1.62x), symbolic-and-code (1.76x), and long-context (2.51x) reasoning on 3 LLMs and 16 datasets, SoT consistently improves mean-baseline accuracy while reducing generated tokens by 62.6% and end-to-end latency by 44.6%. Across 2 VLM scales and 3 reasoning tasks, it improves mean accuracy by 3.8 points over reasoning baselines, with 74.9% fewer completion tokens and 73.5% lower latency than search-based methods. Under constrained access, SoT retains 38.2%/36.5% mean accuracy gains in training-free/embedding-only settings, while trajectory-only judging reaches 84.1% agreement across 3 API models. Together, endogenous state-driven reasoning provides a generalizable and efficient alternative.

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.

Long-term memory is essential for LLM-based agents operating over extended interactions. Existing memory systems primarily update memory when new information arrives, treating retrieval as the endpoint of memory access rather than a driver of memory evolution. Consequently, retrieval feedback is rarely exploited to reorganize memory for future access continuously. Moreover, most existing approaches rely on predefined memory structures together with fixed retrieval pipelines, limiting the agent's ability to organize and evolve its own memory autonomously. Inspired by memory reconsolidation in cognitive neuroscience, we propose REALM, a reconsolidation-evolution agentic long-term memory framework. It models long-term memory as a continual lifecycle by autonomously organizing memories into a heterogeneous cognitive graph, retrieving evidence via adaptively composed graph-search atoms, and continually reconsolidating memories based on retrieval feedback. REALM achieves an average accuracy of 75.97% on LoCoMo and 65.11% on LongMemEval, outperforming the strongest baselines by 7.17 and 1.31 points respectively. Ablation studies confirm that memory reconsolidation consistently boosts performance, with further analyses revealing that it progressively reorganizes related memory units into more coherent local structures for collective evidence recall and utilization during reasoning. These results suggest that retrieval-driven memory reconsolidation provides an effective mechanism for continually evolving long-term memory in LLM agents.