Artificial Intelligence · Cross-listed from cs.LG
Coverage, Not Targeting: A Structural Regime in Multi-Turn Agent Credit Assignment
Chenyu Zhou, Qiliang Jiang, Shuning Wu et al. (+1 more)
Multi-turn agentic RL increasingly treats credit assignment as a targeting problem: given a terminal verifiable reward, per-turn methods localize credit onto the turns that mattered. We identify the structural quantity that predicts when this is the right move, the verifier information density V_d = k/C (the fraction of an
agent's C-step causal chain whose per-turn correctness the verifier exposes), and show that terminal-state verifiers sit deep in a low-V_d regime where targeting is the wrong axis. In controlled shared-rollout comparisons on tau^2-bench that separate reward density from credit geometry, a continuous dense reward spread uniformly beats the sparse binary outcome reward (net-harmful on 4/5 seeds), while concentrating the same advantage on progress turns or on random turns is equally harmful: targeting is second-order. The mechanism is coverage: terminal-state verification collapses the observable signal to a single final-write turn (k=1 in 98% of rollouts) while success requires a 5-8 step chain of prerequisite tool calls. A synthetic phase boundary places the crossover at V_d* ~ 0.8, whereas measured V_d is ~0.15 on tau^2-bench and ~0.4 on BFCL V3; uniform also wins on BFCL, where a matched-concentration shuffled control is negative on 8/8 seeds. The effect reproduces across model families on ToolACE-2-8B (Delta = -0.048 over 32 pre-registered seeds; an independent 20-seed replication is itself significant), and a pre-registered matched-budget breadth sweep traces a monotone dose-response whose deficit vanishes only at full chain coverage, with a reward-to-go arm reaching full-coverage parity. Uniform redistribution is the zero-information coverage default that per-turn schemes must beat; we contribute the matched-concentration shuffled control that any targeting claim should clear.
Read the full abstract ▾
All authors Chenyu Zhou, Qiliang Jiang, Shuning Wu, Xu Zhou
Artificial Intelligence · Cross-listed from cs.CL
Before the Script, Set the Stage: How Worldview Simulation Amplifies Psychologically Grounded Persuasion in Multi-Turn Jailbreaking
Siyu Chen, Haoran Wang, Xiaojian Li et al. (+3 more)
Multi-turn jailbreak attacks demonstrate that harmful intent can be distributed across dialogue, yet existing methods obscure what conversational mechanisms drive vulnerability. We introduce BLUEPRINT, a safety-evaluation framework separating a factorized social-influence strategy space from WORLDVIEWSIM, a cross-turn situational context module. Monte Carlo Tree Search optimizes turn-level combinations of 18 theory-grounded influence factors across a four-turn trajectory. Across six frontier models, BLUEPRINT achieves near-ceiling ASR on major open-weight and proprietary models, while requiring the fewest average queries (2.46). The resulting trajectories further reveal model-specific vulnerability among resistant targets: each responds to distinct influence factors and strategy transitions, yet all share a common recovery pathway-shifting toward concrete, executable task framing consistently escapes hard-refusal states. Ablations confirm operational cues matter most: making requests actionable has the largest impact, gain framing is unusually potent, and some legitimacy appeals can backfire. These findings suggest robust multi-turn safety requires monitoring not only harmful content, but also how dialogue state makes unsafe requests appear concrete and locally executable.
Read the full abstract ▾
All authors Siyu Chen, Haoran Wang, Xiaojian Li, Yao Huang, Yinpeng Dong, Wei Xu
Artificial Intelligence · Cross-listed from cs.LG
Evidence for Shared Routing Geometry and Dynamics in Sparse Mixture-of-Experts
Kirill Labzin, Stepan Kulibaba, Artem Dzhalilov et al. (+1 more)
Sparse mixture-of-experts (MoE) models use an independently parameterized router at each sparse layer to select experts for every
token. Prior work has shown that routing decisions across depth can often be predicted from earlier routing signals, suggesting that routing is not fully independent across layers. However, the structure behind this predictability remains unclear. In this work, we provide evidence that routing-relevant states across layers share a common
geometric structure that is obscured by layer-specific coordinate systems. We isolate the control subspace of each router and align these spaces into a shared canonical representation using generalized orthogonal Procrustes analysis. After
alignment, a single linear transition reaches $R^2=0.39$–$0.71$ and retains 79–90% of the predictive power of separately fitted layer-specific dynamics, indicating that much of routing-state
evolution follows a reusable process across depth. We then ask whether this shared dynamics is specific to routing or simply reflects the smooth
evolution of hidden representations. A matched-rank comparison shows that residual representations are often easier to predict across layers, while router-control states preserve the model's expert choices much more faithfully. This separates generic cross-layer predictability from routing-specific information. Finally, we test whether the predicted canonical states remain meaningful when used in place of native routing states. The transported states preserve local routing behavior, while learned state
evolution reduces $\Delta\mathrm{NLL}$ relative to simple persistence by 15.7% on OLMoE and 6.2% over a 10-router horizon on Phi.
Read the full abstract ▾
All authors Kirill Labzin, Stepan Kulibaba, Artem Dzhalilov, Artem Gorokhov
Artificial Intelligence
Contrastive Explanations in Quantitative Bipolar Argumentation Frameworks
Xiang Yin, Nico Potyka, Antonio Rago et al. (+1 more)
Argumentation frameworks are useful tools for representing and
reasoning with information in a variety of settings, e.g. in supplementing AI models as they perform classification tasks, with a notable benefit of providing additional explainability. In this paper, we introduce contrastive explanations for Quantitative Bipolar Argumentation Frameworks (QBAFs), one such formalism. Unlike most existing explanations for QBAFs, which explain the
reasoning outcome of a single argument of interest (i.e. a topic argument), contrastive explanations explain the difference between two topic arguments. We introduce a general form of contrastive attribution functions (CAFs) and establish a set of general properties they should satisfy. We introduce CAFs based on removal, gradients and Shapley-values, and study their properties. Finally, to illustrate contrastive explanations, we demonstrate their usefulness in healthcare and bias identification settings.
Read the full abstract ▾
All authors Xiang Yin, Nico Potyka, Antonio Rago, Francesca Toni
Artificial Intelligence · Cross-listed from cs.CL
PolERo: Studying Political Evasion in Romanian
Gabriel Stefan and Sergiu Nisioi
Political evasion refers to responses that engage with a question while withholding the requested information. Recent NLP work frames political evasion as a classification task using a two-level taxonomy of response clarity and fine-grained evasion strategies. Existing work on response clarity and evasion classification is limited to English, leaving open whether the taxonomy and model behavior transfer across
languages and political contexts. We introduce PolERo, a
dataset of 3,574 human-annotated question-answer pairs extracted from official transcripts of five Romanian presidents. We evaluate multiple classification approaches on both
datasets under matched conditions, including TF-IDF baselines, fine-tuned encoder models, a proposed sliding-window encoder, and zero/few-shot LLM prompting. We study cross-lingual transfer through joint bilingual training and machine-translation-based data augmentation. Our results indicate that fine-tuned encoders are competitive, cross-lingual transfer is asymmetric, and ambivalent evasion categories involving pragmatic cues remain the main challenge across all model families.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.CL
MultiGhostBench: A Multilingual Benchmark for Long-Form LLM-Generated Text Attribution under Distribution Shifts
Matteo Greco, Anudeex Shetty, Andrea Tagarelli et al. (+1 more)
While existing work on LLM authorship attribution (AA) has made progress, available
benchmarks remain limited, often focusing on English, controlled settings, or relatively outdated models, with the few
multilingual studies considering only relatively short texts. We introduce MultiGhostBench, a
multilingual benchmark comprising 928 books generated by five recent LLMs across six
languages and three scripts, with an average length of approximately 59K words per book. The
benchmark supports evaluation under domain, author, and
language shifts. Evaluation of representative AA methods shows that no single method consistently performs best across settings, and performance generally degrades under distribution shifts.
Transformer-based detectors can retain generator-related information across
languages, although transfer effectiveness varies by
language pair, whereas statistical and fingerprint-based detectors are more
language-dependent. We envision MultiGhostBench as a valuable resource for the development and evaluation of robust AA methods. The
dataset and code can be found at https://github.com/GrecoMT/MultiGhostBench.
Read the full abstract ▾
All authors Matteo Greco, Anudeex Shetty, Andrea Tagarelli, Jey Han Lau
Artificial Intelligence · Cross-listed from cs.LG
Percolation Dynamics in Optimization: Variance Cascades and Nested Symmetry
Sai Niranjan Ramachandran and Suvrit Sra
We study the dynamics of
Stochastic Gradient Descent (SGD), which is known to steer deep neural networks toward invariant sets that correspond to simpler subnetworks. How this steering unfolds over time remains poorly understood. We answer this by modeling the
stochastic gradient flow (SGF) as a percolation process, in which nested architectural
symmetries force subnetworks to merge in discrete blocks rather than by single-edge attachment. These structural transitions register as variance spikes in a macroscopic order parameter echoing physical phase transitions. We further state sufficient conditions under which the trapping argument carries over to Adam and AdamW under heavy-tailed gradient noise and measure them on a trained
Transformer.
Read the full abstract ▾
Artificial Intelligence
Diagnosing with Insights: Structured Analysis of Agent Failures via Behavioral Abstractions
Jiayi Bi, Yanjie Gao, Yuanmin Xie et al. (+4 more)
With the proliferation of LLM
agents, the ability to understand and diagnose failures in
agents is essential to achieving superior effectiveness and trustworthiness. As
agent failures often manifest via long and complex trajectories, manually finding the needles in the haystack is untenable. However, traditional diagnosis techniques for software bugs can hardly address LLM
agent failures, while completely relying on LLMs as the judge yields unreliable diagnosis results. To overcome these challenges, this paper presents AGENTSCOPE, a new neuro-symbolic approach for
agent failure mode diagnosis. The key principle of AGENTSCOPE is to abstract
agent behavior, based on its trajectories, into structured representations. Furthermore, AGENTSCOPE introduces the concept of neural invariants to specify
agent behavior properties. AGENTSCOPE leverages LLM-guided
reasoning atop the structured representation against neural invariants to pinpoint both the failure step and its type in the trajectory. We show the effectiveness of AGENTSCOPE on publicly available
agent failure
datasets (Who&When) and a more comprehensive
dataset created by us (AgentErrata), where AGENTSCOPE significantly outperforms the current state of the art in fault localization and attribution accuracy. Our work shows that integrating structured abstractions with LLM-guided
reasoning enables effective, reliable, and interpretable diagnosis for
agent failures.
Read the full abstract ▾
All authors Jiayi Bi, Yanjie Gao, Yuanmin Xie, Liqun Li, Tianyin Xu, Fan Yang, Mao Yang
Artificial Intelligence · Cross-listed from cs.CL
NE-R1: Enhancing Named Entity Recognition Model via Reinforcement Learning
Meixuan Chen, Hehan Li, Ruizhi Zhao et al. (+8 more)
Named Entity Recognition (NER) has achieved substantial progress since the advent of large
language models (LLMs). Nevertheless, the recognition of long-tail and domain-specific entities remains challenging due to the deficiency in parametric knowledge.
Retrieval-augmented generation (RAG) offers a promising remedy by injecting external knowledge, but it also introduces noise and unnecessary cost when dealing with familiar cases. In this paper, we propose NE-R1, a novel framework for adaptive
retrieval-augmented NER. We design a "
retrieval-on-demand" mechanism for NER. Then we integrate it into models by a two-stage training method: (1) multi-task instruction tuning initialization; (2) end-to-end RL optimization with CoT. To achieve reasonable selection between parameterized and external knowledge, we design a multi-dimensional reward considering both accuracy and
retrieval benefit. NE-R1 achieves state-of-the-art performance on various
benchmarks, with an average F1 score gain of 2.52% in in-domain evaluation and 1.18% in zero-shot cross-domain evaluation.
Read the full abstract ▾
All authors Meixuan Chen, Hehan Li, Ruizhi Zhao, Xin Lu, peizhi xu, Liwei Qian, LI Meifang, shuanglong li, Hanmeng Liu, Xin Pei, Yanbiao Ma
Artificial Intelligence · Cross-listed from cs.HC
Towards a Foundational Ontology for Identifying and Resolving Contradictions in Dialogue-based Human-Robot Interactions
Maitreyee Tewari and Michele Persiani
Existing Human-
Robot Interaction (HRI) literature has focused on identifying and structuring errors, failures, conflicts, and knowledge issues (called in this work as contradictions) in domain-specific dialogue-based interactions. However, there is still lack of a formal computational framework to represent and define these contradictions, interoperable and usable across HRI and human-
agent interaction (HAI) domains. Thus, this research project aims to capture, represent, and evaluate the notion of (1) dialogue-based collaborative interaction and (2) related contradictions in a foundational ontology. METHONTOLOGY, a systematic approach to build domain-independent ontologies was applied. In the conceptualisation stage of the presented ontology, concepts and models from Activity Theory were used. Preliminary results presented in this short article are: (i) Natural
language definitions of dialogues and related contradictions in HRI, (ii) Set Theoretic definitions of dialogues and contradictions, and (iii) First Order Logic (FoL) formulation of the contradiction concepts and three novel principles guiding dialogue-based interactions between humans and
robots. In summary, we report on ongoing work to develop a foundational ontology based on Activity Theory called Activity Theory-based foundational ontology (ATFOt) to capture and represent the notion of contradictions in HRI.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.GT
Fair Stable Matching: A Nash Social Welfare Approach
Parth Desai, Rasheed M, Ganesh Ghalme et al. (+1 more)
While traditional stable matching
algorithms, such as the Gale-Shapley
algorithm, prioritize stability, they may fall short of achieving equitable outcomes among participants. We study the role of Nash social welfare (NSW) as a
fairness objective in the classic stable marriage problem. We develop SNSW-Alg that finds a stable matching that maximizes Nash social welfare under rank-induced utilities in $\tilde{\mathcal{O}}(n^4)$ time, where $n$ is the number of men or women. We demonstrate that SNSW-Alg balances equity while preserving stability. We empirically evaluate our methods across diverse preference distributions, demonstrating significant gains in
fairness without substantial losses in other key measures such as regret, egalitarian criterion, and sex equality. Our findings suggest that the stable matching produced by SNSW-Alg is statistically Pareto-undominated by stable matchings based on other
fairness measures - regret, egalitarian, and sex equality. This study offers compelling insights for designing fair-stable matching.
Read the full abstract ▾
All authors Parth Desai, Rasheed M, Ganesh Ghalme, Sujit Gujar
Artificial Intelligence · Cross-listed from q-bio.GN
Subcellularly Resolved Single-Cell Embedding Learning with Transcriptomic data, Protein Structure and Localization Information
Zhen Zhou, Jiachen Li, Yuan Liu et al. (+2 more)
Existing
cell embedding methods predominantly rely on transcriptomic or proteomic measurements and represent each
cell as a holistic entity, thereby overlooking the subcellular localization of individual molecules. Moreover, they rarely incorporate
protein structural information, despite its fundamental role in determining molecular interactions and functions. In this work, we propose a
multimodal framework for learning subcellularly resolved
cell embeddings by jointly leveraging RNA expression profiles,
protein sequence representations, and
protein structural information. Specifically, we employ a cross-
attention architecture to integrate transcriptomic, sequence, and structural modalities and model their interactions within distinct subcellular compartments. The resulting
embeddings represent each
cell through its fine-grained subcellular organization, capturing both molecular expression patterns and the functional properties of the associated
proteins. By learning
cell representations at subcellular resolution, our framework preserves spatially organized biological information while integrating complementary signals across multiple molecular levels. To the best of our knowledge, this is the first framework that produces subcellularly resolved
cell embeddings by jointly incorporating transcriptomic information,
protein sequence representations, and
protein structural knowledge within a unified cross-modal learning paradigm.
Read the full abstract ▾
All authors Zhen Zhou, Jiachen Li, Yuan Liu, Xiaoyong Pan, Hong-Bin Shen
Artificial Intelligence · Cross-listed from cs.LG
AGI Maze Prediction Datasets: A Compact Benchmark for Learning World Dynamics with Transformers
Alexey Potapov
World modeling requires a predictive model to maintain and update an internal state adequate for
reasoning about the consequences of actions. We introduce the AGI Maze Prediction
Datasets and
Benchmark, a lightweight controlled testbed for studying this capability in
Transformers and other predictive models. Derived from procedurally generated, stateful grid worlds, the
benchmark comprises per-step transition prediction, fixed-horizon state prediction, and sequential textual-observation prediction. Source-maze-disjoint training and validation splits, together with greedy exact-match evaluation, distinguish learning transferable action-conditioned dynamics from memorizing transitions in familiar layouts. We establish from-scratch byte-level
Transformer baselines and compare them with two working-
memory-augmented architectures. A generic auxiliary latent-
memory Transformer can fit some training sets perfectly but does not consistently improve held-out performance. In contrast, a pseudo-
video spatial-
memory Transformer initializes a two-dimensional latent workspace from the input map and updates it from action history without receiving intermediate maps, positions, or state labels. Under the same data, objectives, and evaluation protocol, this model reaches perfect validation accuracy on selected fixed-horizon tasks where the byte and unstructured-
memory baselines do not, and substantially improves sequential text-trace prediction. These results suggest that structured, task-aligned working
memory can be more useful than additional latent capacity alone. More broadly, we argue that
language grounding is mediated by persistent data structures and computations over them; the
benchmark offers a compact setting for testing architectures that couple textual interfaces to learned structured state.
Read the full abstract ▾
Artificial Intelligence
SALA: Semantic-Aware Logical Alignment for Complex Reasoning in In-Context Learning
Zhao Ji, Wenqing Chen, Zhixuan Chu et al. (+4 more)
Effective in-context learning (ICL) for complex
reasoning relies on selecting the right demonstrations. Traditional
retrieval methods based on surface similarity fail to capture the underlying problem-solving logic. Recent logic-based methods address this by matching predefined
reasoning steps, but the rigid rules and exact-match criteria is improper to handle flexible or diverse
reasoning processes. To address the problem, we propose SALA, a Semantic-Aware Logical
Alignment framework. Instead of relying on a fixed inventory, SALA automatically learns task-specific
reasoning operations. It then embeds these operations into a continuous semantic space and uses dynamic time warping (DTW) to align the
reasoning sequences. This approach allows for soft, flexible matching of
reasoning logic while remaining highly interpretable. Experiments across four
reasoning benchmarks and three LLMs demonstrate that SALA outperforms existing demonstration selection methods. Further analysis confirms the roles of the operation induction and the logical semantic
alignment.
Read the full abstract ▾
All authors Zhao Ji, Wenqing Chen, Zhixuan Chu, Jianxing Yu, Jingping Liu, Shanhe Zhao, Zibin Zheng
Artificial Intelligence · Cross-listed from cs.CV
ORB-SVM : An Innovative Hybrid Framework for Efficient Brain Tumor Detection from MRI Scans
Amirhosein Azarpour
Brain cancer remains one of the most significant challenges in modern medicine, where the accuracy of early stage diagnosis is a decisive factor in patient survival and treatment efficacy. Although
Magnetic Resonance Imaging (MRI) is the established gold standard for visualizing neurological structures, the interpretation of these high dimensional scans is often complicated by subjective variability among practitioners and the inherent noise present in complex medical images. While contemporary approaches frequently rely on high parameter deep learning architectures, such models often involve significant computational costs and require extensive data for effective training. This study introduces a hybrid framework that utilizes the Oriented FAST and Rotated BRIEF (ORB)
algorithm for precise feature extraction and a Support Vector Machine (SVM) for classification [1], [2]. The proposed approach achieves a sub- stantial data reduction of approximately 99.5%, which effectively minimizes the influence of non informative background data while preserving critical diagnostic patterns essential for tumor identification. By balancing feature sparsity with a robust kernel based classifier, this methodology addresses the limitations of over parameterized systems while maintaining high diagnostic integrity. Experimental evaluations conducted on the Br35H
dataset demonstrate that the framework attains a classification accuracy of 97.5%. The findings suggest that the integration of localized feature representation and optimized classification provides a reliable and resource efficient alternative for medical image analysis, offering a structured solution that maintains per- formance without the need for extensive computational overhead.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.LG
What Is Worth Representing? Representational Empowerment for Continual Model Construction
Fei Dai, Hanqi Zhou, Alison Gopnik et al. (+1 more)
The first problem of modeling the world is not just estimating the right parameters or causal structure, but deciding what should be represented at all. We frame this problem as continual model construction: an
agent maintains an environment-specific model M of an inaccessible world W and curates a persistent library L of reusable representational elements across environments. We propose Representational Empowerment (RepEmp) to score candidate elements by how much they expand the
agent's future capacity to model and plan, complementing the classic definition of empowerment, but redefined as control over internal representations instead of external states. We realize the framework as a hierarchical Curator-Actor architecture and test it across three experiments. In a closed-vocabulary causal-learning task, human participants construct causal models at varying abstraction granularities to maximize goal reachability rather than fidelity to the world, a signature better predicted by RepEmp than by information-gain alternatives. Matched simulations reveal that RepEmp-guided construction contributes more than exploration to sufficient structure recovery and cross-task transfer. Finally, in an open-vocabulary planning domain, an LLM-augmented Curator builds more compact symbolic libraries, which also generalize better than baselines. Ablating RepEmp eliminates these benefits. Together, these results identify RepEmp as a key principle for continual model construction: deciding what to build, retain, and reuse under bounded resources.
Read the full abstract ▾
All authors Fei Dai, Hanqi Zhou, Alison Gopnik, Charley Wu
Artificial Intelligence · Cross-listed from cs.CL
DiffIE: Diffusion-based Open Information Extraction
Konstantin Fedorov and Valentin Malykh
A single sentence often expresses multiple valid relational triplets, which makes Open Information Extraction (OpenIE) fundamentally a multi-output task. Existing neural systems handle this by autoregressive generation, which is flexible but slow and prone to redundancy, or by fixed-slot prediction, which is efficient but couples the extraction budget to training. We introduce DIFFIE which instead treats the stochasticity of conditional discrete
diffusion as the extraction mechanism itself: independent reverse-
diffusion trajectories over per-
token role tags produce a pool of candidate triplets, which are clustered under lenient matching and ranked to form the output. Both the pool size and the number of returned extractions are inference-time choices, decoupling the extraction budget from training and exposing test-time compute as a tunable axis. DIFFIE achieves the new state of the art in CaRB (1-1) both F1 and AUC, and outperforms the strongest rule-based system (ClausIE) in BenchIE; it also remains competitive in standard CaRB and WiRe57 evaluations, giving the best average score among systems that report all four
benchmarks. Ablations show that uniform discrete
diffusion outperforms absorbing state
diffusion in our setting, and that a matched non-
diffusion stochastic tagger does not reproduce its gains. Our results indicate that
diffusion stochasticity is an effective mechanism for structured prediction tasks with multiple valid outputs.
Read the full abstract ▾
Artificial Intelligence
Improving Evaluation Realism with Inference-Time Compute and Deployment Scaffolds
Axel Ahlqvist, Richard Guan, Juan-Pablo Rivera et al. (+6 more)
A core obstacle to
alignment evaluation is evaluation awareness: capable models can tell when they are being tested rather than deployed, weakening the conclusions a safety evaluation can support. We present two techniques that make simulated
alignment evaluations harder to distinguish from real deployments. Our first technique, critique refinement, spends additional inference-time compute on each simulator action: the simulator generates multiple candidate actions, refines them using feedback from an instance of the target model on how to make them more realistic, and continues the evaluation with the most deployment-like candidate. Our second technique, DISH (Deployment-Imitating SWE-
Agent Harness), wraps the target in an
agent harness, reducing the gap between simulated and real deployment environments in coding settings. We test the techniques on multiple target models and find that they compose: applying both yields larger realism gains than either alone. Our results show that automated approaches can improve the realism of
alignment evaluations, and that these improvements use additional compute more effectively than making the audits longer.
Read the full abstract ▾
All authors Axel Ahlqvist, Richard Guan, Juan-Pablo Rivera, Adeline Kassler, Dmitrii Troitskii, Alexandra Souly, Kai Fronsdal, Robert Kirk, John Hughes
Artificial Intelligence · Cross-listed from cs.LG
SEAL: Reinforcing Global Safety in Mixture-of-Experts through Shared Expert ALignment
Qingyu Meng, Yiwei Zha, Jiahuan Pei et al. (+3 more)
Mixture-of-Experts (MoE) is a scaling architecture for large
language models that activates only a small subset of expert modules per
token, enabling massive parameter growth with nearly constant computation. Recent Hybrid MoE architecture adds shared experts to capture consistently useful representations, further improving stability and generalization. MoE now powers many flagship open-source and commercial models, yet remains vulnerable to
adversarial attacks. Specifically, sparse routing introduces a structural vulnerability: MoE safety hinges on which experts are activated, and adversaries can subvert this selection through jailbreak
prompts, malicious fine-tuning, and weight-level
pruning of safety-critical
neurons. Existing defenses primarily focus on hardening the router, but an adversary may still manipulate or bypass the routing trajectory due to the routing process's nondeterministic nature, thereby collapsing the defense. To cope with this problem, we first identify theoretically and empirically that shared expert, an always-activated component containing a small proportion of safety-critical
neurons, can overcome the uncertainty of sparsely activated routing path and serve as a router-independent anchor to enhance global safety
alignment. Based on this insight, we propose SEAL, a training-time parameter-efficient defense that produces a plug-and-play adapter attached to shared expert, and SEAL++, a variant that adds an orthogonal constraint preserving pre-existing safety subspaces during training. We evaluate SEAL and SEAL++ across six attack scenarios that combine three
adversarial inputs (harmful prompting, jailbreak, malicious fine-tuning) with and without
neuron pruning. SEAL reduces attack success rate (ASR) by up to 60%, at a capability cost of at most 1.4% on a five-
benchmark average. Additionally, SEAL can seamlessly integrate with router-level ......
Read the full abstract ▾
All authors Qingyu Meng, Yiwei Zha, Jiahuan Pei, Koen Hindriks, Herbert Bos, Min Chen
Artificial Intelligence
SCX Router: Streaming Zero-Shot Model Selection with a Decoder-KV Classifier and a Real-World Task Ontology
Ihor Stepanov, Aleksandr Smechov, Mykhailo Shtopko et al. (+2 more)
The rapid proliferation of large
language models (LLMs) and the growing diversity of their applications presents a unique optimization opportunity: selecting the right model for the task, while optimizing for speed, cost, and quality at a per-task level. However, inference endpoints can vary widely in quality, price, latency, context support, tool use, domain expertise, and
reasoning behavior. This heterogeneity makes manual heuristics difficult to maintain and unlikely to achieve consistently favorable speed–cost–quality trade-offs on their own. We introduce \router{}, a lightweight GLiClass-based router that assigns a suitability score to each inference-time model label without autoregressive generation. The released 0.6B-parameter checkpoint combines a Qwen3 decoder with a shallow bidirectional scorer. Its decoder-KV execution path preserves a text-only key–value cache across a session, encodes only new dialogue turns, and evaluates transient candidate-label
tokens without adding them to the persistent cache. The same checkpoint also predicts task type, difficulty,
reasoning mode, and expected output length, and supports custom zero-shot labels. For task generation, we construct a task ontology with 23 families, 115 task types, 345 routable subtypes, 1,173 synthetic examples, and an orthogonal axis of 30 domains. Using this structure, we generate 150,000 verifier-scored tasks and 15,000 open-ended tasks. We then train the Qwen3 decoder on these tasks, while explicitly separating learned request prediction from per-task policies for attributes such as eligibility, cost, cache reuse, safety, and sovereignty. Across six LiveBench subsets, the router outperforms the mean candidate; on the selected 1,000-task subset, it achieves an aggregate top-1 score of 0.707 versus 0.696 for the strongest fixed model, with
benchmark-dependent gains.
Read the full abstract ▾
All authors Ihor Stepanov, Aleksandr Smechov, Mykhailo Shtopko, Dmytro Vodianytskyi, Oleksandr Lukashov
Artificial Intelligence · Cross-listed from cs.CV
VoRTeC: Taming Foundation Flow for One-step Real time Video Compression
Yichong Xia, Qinhong Wu, Bin Chen et al. (+3 more)
Ultra-low bitrate
video compression still faces critical challenges: traditional neural
video compression inevitably introduces blurring artifacts, while
diffusion-based generative
video compression suffers from excessive decoding latency and poor temporal consistency. To address these issues, we propose $\mathtt{VoRTeC}$, a
Video Compression framework built upon a foundational flow model (Wan2.1). By compactly encoding latent
video representations, predicting the positions of compressed representations along flow trajectories, and integrating multi-scale priors, $\mathtt{VoRTeC}$ enables the compressor to harness generative
video flow priors effectively. Without accessing the parameters or gradients of flow matching networks, our framework achieves one-step decoding and reconstructions with high perceptual fidelity. Meanwhile, we maintain consistency across frame groups via tail-frame reuse and prior caching. Extensive experiments demonstrate that our method reduces bit consumption by 58% compared to prior
diffusion-based approaches, with decoding speed boosted by 3 to 197 times: $\mathtt{VoRTeC}$ achieves a decoding speed of 13 FPS at 720p and 32 FPS at 480p.
Read the full abstract ▾
All authors Yichong Xia, Qinhong Wu, Bin Chen, Jinpeng Wang, Zeyuan Chen, Haoqian Wang
Artificial Intelligence · Cross-listed from cs.CV
RouteGraph-Mona: Confusion-Aware Routing Fine-Tuning for Mineral Image Classification
Jierui Li, Zhiyuan Qi, Hao Zhu et al. (+7 more)
Mineral image classification is important for geological exploration and resource development, but it remains challenging due to substantial intra-class variations in appearance and high inter-class visual similarity. Multi-cognitive Visual Adapter (Mona) is a
vision-oriented parameter-efficient adapter that adapts pre-trained visual models by tuning only a few parameters. However, Mona statically aggregates responses from multiple scales, limiting its ability to accommodate sample-specific scale preferences and model confusion among visually similar mineral categories. To address this issue, we propose RouteGraph-Mona, a lightweight route-space regularization method built on Mona. Specifically, we replace Mona's static multi-scale aggregation with sample-adaptive routing. The resulting branch-selection behavior defines a compact routing space that captures each image's scale preferences. We then regularize the resulting routing signatures with class-wise route anchors and confusion-weighted margins. The route anchors encourage class-consistent routing patterns, while the margins promote greater separation between visually similar categories in the routing space. Experiments on three public mineral image
datasets with two visual backbones show that RouteGraph-Mona consistently outperforms Mona in mean accuracy and remains competitive with representative fine-tuning methods and mineral image classification baselines.
Read the full abstract ▾
All authors Jierui Li, Zhiyuan Qi, Hao Zhu, Yufan Liu, Jixian Liu, Shaojie Jiang, Jianda Wang, Yaqi Liu, Xiaotong Li, Wei Wang
Artificial Intelligence · Cross-listed from cs.SD
Auditory Illusion Benchmark for Large Audio Language Models
Hayoon Kim, Eunice Hong and Kyogu Lee
Perceptual illusions have long served as crucial probes into human cognition, revealing biases and limitations of perception. In the auditory domain, such illusions provide a unique lens for testing whether Large Audio
Language Models (LALMs) replicate human perceptual tendencies. Despite their importance, most
benchmarks focus on visual illusions or general audio tasks, leaving auditory illusions underexplored. To this end, we present AIB, the first auditory illusion
benchmark for LALMs, covering ten representative illusions across music, sound, and
speech, each annotated for the presence of knowledge-based priors. Our methodology pairs model evaluation with controlled human listening studies, enabling direct comparison of responses. Results show systematic differences: while most LALMs remain signal-faithful on low-level acoustic illusions, several exhibit more human-like responses when linguistic or musical priors are involved, although no model matches the human perceptual profile. These findings highlight the current limitations of LALMs as cognitive models. By establishing auditory illusions as a rigorous testbed, our work offers a new perspective for probing neural black-box models and advancing understanding of auditory cognition. AIB is publicly available at https://github.com/gillosae/aib.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.CL
Do Large Language Models Capture the Diversity in their Training Data?
Youqi Wu and Farzan Farnia
Large
language models are trained to model conditional distributions over text, yet it remains inadequately understood whether they capture the full diversity of plausible outputs present in their training data. We study this question through an information-theoretic lens by comparing the conditional
entropy of model-generated outputs with that of the corresponding training data. Given paired input-output samples, we use conditional
entropy and its
matrix-based analogue based on von Neumann
entropy to measure output variability beyond what is explained by the conditioning input, without requiring multiple reference outputs for the same
prompt. Across LLM families with publicly available training data, including OLMo, Pythia, and GPT-Neo, we consistently find that model-generated outputs exhibit lower conditional
entropy than their training data, across different model scales, sequence lengths, and decoding strategies. We observe a similar conditional diversity gap beyond
language modeling, including class-conditioned ImageNet generators and text-conditioned models trained on MS-COCO. To address this gap, we propose a post-hoc correction mechanism that generates multiple outputs for each input and reweights them through a
matrix-
entropy projection, increasing conditional diversity while remaining close to the original model distribution. We prove the concavity of the
matrix-based conditional
entropy functional, which makes the resulting
entropy-constrained projection a convex optimization problem, and develop a scalable mirror-descent
algorithm for its implementation. Our results reveal a systematic conditional diversity gap between modern generative models and their training data, and provide an information-theoretic framework for measuring and mitigating this gap.
Read the full abstract ▾
Artificial Intelligence
CoMerge: Conflict-Driven Preference Optimization for Multi-Task Model Merging
Mingjie Zheng, Zihao Chen, Wenqing Chen et al. (+4 more)
Model merging provides an efficient paradigm for constructing multi-task large
language models (LLMs) without full model retraining, yet it remains challenged by parameter interference. While existing methods aim to preserve the capabilities of individual expert models and mitigate interference, they generally do not directly learn from the potentially degraded behaviors exposed by naive merging. In this paper, we propose a conflict-driven preference optimization framework for model merging (CoMerge), which reformulates model merging as a preference optimization problem. The approach utilizes a self-supervised, conflict-driven strategy that leverages the defects of naive merging methods (e.g., task arithmetic) as hard negative samples to construct preference pairs without external annotations. By applying preference optimization to refine lightweight, tensor-wise merging coefficients, CoMerge enables the model to mitigate parameter-space conflicts while preserving task-specific capabilities. Extensive experiments show that CoMerge achieves an average normalized performance of 0.9968 on MergeBench, outperforming all evaluated data-free and data-driven model-merging baselines. Furthermore, on Llama-3.1-8B-Instruct, CoMerge yields marked improvements on conflict-sensitive tasks such as instruction following and safety, while remaining highly competitive with full-parameter fine-tuning despite optimizing only 1,445 scalar coefficients.
Read the full abstract ▾
All authors Mingjie Zheng, Zihao Chen, Wenqing Chen, Weile Yuan, Zhixuan Chu, Jianxing Yu, Zibin Zheng
Artificial Intelligence · Cross-listed from cs.CL
PaperCompiler: Faithful Paper-to-Code Generation via Repository-Level Specification Compilation
Yunhao Liu, Hong Phuc Pham and Jaehong Yoon
Faithfully translating research papers into repository-level implementations remains challenging because papers often describe methods at a high level, leave implementation assumptions implicit, and require generated repositories to preserve method logic, evaluation protocols, and cross-file consistency. Despite recent advances in paper-to-code
agents, their intermediate outputs are often presented as free-form plans or summaries that downstream coding
agents may ignore, reinterpret, or compress, leading to algorithmic simplification and inconsistent repository structure. To address these challenges, we introduce PaperCompiler, a paper-to-code generation framework that compiles paper-grounded evidence into explicit repository-level implementation specifications. PaperCompiler grounds implementation-relevant evidence while preserving source provenance and distinguishing paper-supported, inferred, externally delegated, and unresolved information. The resulting specifications encode non-degradation requirements, ownership assignments, cross-file dependencies, and file-level constraints. Repository generation proceeds under these compiled specifications while retaining flexibility over local engineering choices not fixed by the paper. PaperCompiler outperforms strong baselines on Paper2CodeBench, achieving a 13.8% relative improvement in reference-based fidelity (from 3.64 to 4.15) and reducing high-severity evaluator critiques (from 13.2% to 6.1%).
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.RO
CrashDiffuser: VLM-Guided Collision Intent Reasoning for Fine-Grained Safety-Critical Traffic Scenario Generation
Shucheng Zhang, Yuang Zhang, Bingzhang Wang et al. (+3 more)
Generating safety-critical scenarios is essential for evaluating autonomous driving systems. However, existing generators primarily focus on inducing collisions and offer limited control over where contact occurs on the target vehicle. In this paper, we study fine-grained safety-critical scenario generation, where success requires both a target collision and a specified head, rear, or side contact region. We propose CrashDiffuser, a closed-loop VLM-guided
diffusion framework that decouples semantic collision
reasoning from continuous trajectory synthesis through a hierarchical collision-intent interface derived from the requested target contact region. At initialization, the VLM extracts reusable scene-level context; at each replanning step, it predicts a structured action tuple describing speed change, turning behavior, and collision stage. This intent conditions a
diffusion model to generate executable
adversarial trajectories, while collision-guided sampling, candidate selection, and short-horizon replanning adapt generation to the target vehicle's evolving behavior. On WOMD-derived closed-loop scenarios, CrashDiffuser achieves a target-collision rate of 50.33% in a single attempt and 67.98% after three attempts, together with a contact-region control success rate of 40.05% and competitive trajectory naturalness. Component ablations further support the proposed design.
Read the full abstract ▾
All authors Shucheng Zhang, Yuang Zhang, Bingzhang Wang, Muhammad Monjurul Karim, Kehua Chen, Yinhai Wang
Artificial Intelligence · Cross-listed from cs.CR
Retrosynthesis of Synthetic Media for Explainable AI Provenance Forensics
Yijie Lin, Ching-Chun Chang, Isao Echizen et al. (+2 more)
With the rapid proliferation of generative models on Machine Learning as a Service (MLaaS) platforms, reliably tracing the provenance of synthetic media without modifying generator architectures or parameters remains a major challenge. In this work, we propose a self-referential retrosynthesis framework for explainable AI provenance forensics under a fixed-generator setting. The framework leverages a jointly optimized encoder-decoder pair to implement a self-
embedding mechanism that enables round-trip consistency verification. During inference, client inputs are first encoded and then processed by the generator to produce outputs with high visual fidelity. For forensic verification, the consistency between the resynthesized image and the query image is analyzed to determine whether the image originates from the target generative model. Our approach eliminates the need for watermark
embedding or modifications to the generation process. Experimental results show that images generated from encoded inputs maintain visual quality comparable to original generator outputs, while decoded images reliably trace back to their corresponding source inputs. Furthermore, the framework provides interpretable evidence for generative content provenance, establishing a practical tool for explainable generative AI forensics.
Read the full abstract ▾
All authors Yijie Lin, Ching-Chun Chang, Isao Echizen, Hui Li, Chin-Chen Chang
Artificial Intelligence
Codebook Agent: Amortized Topology Design for LLM Multi-Agent Systems
Jinxi Yu, Yubei Li, Eric Hanchen Jiang et al. (+6 more)
Adapting the communication
topology of an LLM multi-
agent system to each query improves both accuracy and efficiency, yet current designers treat this as conditional graph generation: a variational, autoregressive, or
diffusion decoder searches the $N \times N$ adjacency space, and a graph-network proxy trained on utility and a structural cost such as edge count ranks the sampled candidates. We argue that this formulation is misaligned with the problem. Empirically,
topologies that survive a reward filter collapse to about six distinct graphs even when the codebook capacity grows from 8 to 64; edge count is negatively correlated with measured
token consumption (Pearson $r \approx -0.4$), so sparsifying the graph makes inference more expensive; and a message-passing scorer over
agent-profile nodes is adjacency-invariant whenever
agents share a profile—the default configuration of published
benchmarks—so it cannot rank candidates at all in that regime. These three facts motivate Codebook
Agent: a vector-quantized autoencoder compresses successful
topologies into a query-independent 16-entry codebook; a reward-weighted MLP maps the query
embedding to a distribution over codes; and an MLP proxy that reads the flattened adjacency, regressed on measured utility and per-task normalized
token cost, reranks the top decoded candidates in a single batched forward pass. With no iterative search and no message passing at test time, Codebook
Agent is the most accurate method on all six
benchmarks we compare (84.6 average against 83.0 for the strongest prior designer), emits a
topology in 2.4 ms, and uses 21.9–33.2% fewer LLM
tokens.
Read the full abstract ▾
All authors Jinxi Yu, Yubei Li, Eric Hanchen Jiang, Zhi Zhang, Dong Liu, Wenxiao Zhao, Levina Li, Kai-Wei Chang, Ying Nian Wu
Artificial Intelligence
APEx: Distillation of Agent Procedural Experience for Adaptive Deep Research Question Answering
Jie Ding, Rui Sun, Xinyuan Zhang et al. (+2 more)
Deep research
agents augment large
language models with external tools to answer complex, long-horizon questions through multi-turn
reasoning. Learning from prior experience is crucial for continual improvement, yet existing methods either retrieve verbose task-specific traces that burden decision-making, or distill procedural skills that remain decoupled from downstream policy adaptation. We propose APEx, a hierarchical experience utilization framework that organizes interaction history into instance-level trajectory
memories and category-level procedural skills, and couples them through a closed-loop architecture of Executor, Distiller, and Planner. The three modules are optimized via a three-stage alternating GRPO training paradigm, enabling reward-guided skill
distillation rather than fixed-
prompt generation. At test time, distilled skills serve as procedural priors for online Planner adaptation through skill-guided test-time reinforcement learning, allowing ground-truth-free self-improvement with skill-
alignment regularization to prevent policy drift. Experiments on 7
benchmarks demonstrate that APEx achieves state-of-the-art performance, surpassing GPT-5.4 by 14.7 points and the strongest
memory-augmented baseline by 3.0 points.
Read the full abstract ▾
All authors Jie Ding, Rui Sun, Xinyuan Zhang, Zeyu Zhang, Xin Liu