Artificial Intelligence · Cross-listed from cs.LG
FinVector-Market-4B: A Controlled Study of LoRA Adaptation for Structured Financial Tasks
Alina Khaybullina
FinVector-
Market-4B adapts Qwen/Qwen3.5-4B with rank-16 LoRA on a 22,000-example corpus for structured financial tasks. We evaluate the base and adapted models on the same 600-example
benchmark under implicit and explicit JSON-schema contracts. Supplying the schema alone raises base-model JSON validity from 0% to 91.3%. Under matched explicit prompting, the frozen scores improve from 14.7% to 40.0% for FinQA answer exact match, from 48.0% to 82.7% for calculator-expression correctness, from 20.1% to 89.5% for scenario branch-label agreement, and from 52.4% to 87.2% for implication-direction agreement. A post-hoc policy-scoring audit shows that the reported macro-F1 decline reflects a changing label set; using the same three target classes gives 77.4% for the base and 83.1% for the adapter. Filing overlap and calculator-target inconsistencies qualify the
benchmark's generalization claims. The results show that compact financial domain adaptation can produce substantial task-specific gains beyond output-format learning under matched prompting, with gains bounded by the evaluated task distribution and
prompt contract.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.GR
Physics-based Sphere Packing for Lagrangian Mesh Morphing
Jiong Lin and Hod Lipson
This paper studies tetrahedral meshes as the body representation for differentiable simulation and computational design. Fixed-connectivity meshes degrade under large morphs, while remeshing from scratch discards node correspondence. We present JamTet, a physics-based sphere-packing framework for volumetric meshing and morphing. We contribute (i) a GPU-parallel mesher combining octree-hierarchical packing with constrained Delaunay tetrahedralization, producing more uniform element volumes than TetGen and fTetWild; (ii) Lagrangian mesh morphing that preserves interior-node identities by re-equilibrating the same spheres within changing shapes and rebuilding the boundary and connectivity, remaining inversion-free where fixed-connectivity and TetSphere meshes invert; and (iii) a differentiable GPU simulator in JAX, with mass-spring edges and a volumetric Neo-Hookean term, integrated with mesh morphing in a design pipeline. In soft-
robot morphology design experiments, interior-node gradients improve swimming fitness by 0.73-1.07 over a matched surface-only variant, while voxelized versions of the same designs yield 32-63% lower fitness. These results establish sphere packing as a practical volumetric mesh representation for gradient-based shape optimization. Code and media: under review.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.CL
Tiny-Scale Chinese BERT Pretraining: A Controlled Comparison of MLM, WWM, and MacBERT Strategies
Yiping Bai
Pretraining strategies significantly impact the quality of
language models, yet existing comparisons of Masked
Language Modeling (MLM), Whole Word Masking (WWM), and MacBERT-style replacement have focused primarily on base-scale models (>=110M parameters). This paper presents a controlled comparison of these three strategies on a tiny-scale Chinese BERT model (4 layers, 256 hidden dimensions, 8.7M parameters). Under identical architecture, corpus (1.29M sentences from Chinese Wikipedia), and hyperparameters, we train three models from scratch and evaluate them across five intrinsic dimensions: perplexity, MLM hit rate, semantic discrimination, grammatical judgment, and contextual sensitivity. At tiny scale, MLM achieves the best overall intrinsic performance (winning 3 of 5 dimensions), while WWM excels in both perplexity (1.27 vs. 2.10, a 39.5% improvement) and MLM hit rate (22% vs. 16%). Notably, MacBERT under a severely limited synonym dictionary (222 entries, 3.3% coverage) exhibits severe perplexity degradation (47.23, 22x higher than MLM), yielding a ranking (MLM > WWM >> MacBERT) that differs markedly from the established base-scale conclusion (MacBERT > WWM > MLM). We further identify a critical evaluation pitfall: MacBERT achieves the lowest training loss (2.17) yet the highest perplexity (47.23), revealing that training loss alone is unreliable under mixed replacement strategies. All models and corpus are publicly available at https://huggingface.co/eebyp.
Read the full abstract ▾
Artificial Intelligence
How Could AI Eliminate Humanity? A Failure-Mode Analysis of Civilizational Risk
Mikołaj Sienicki and Krzysztof Sienicki
This article develops a failure-mode framework for analyzing how advanced artificial intelligence could contribute to human extinction, irreversible civilizational collapse, or permanent human disempowerment. The central thesis is that catastrophic AI risk does not require consciousness, hostility, or an explicit intention to harm humanity. Instead, risk may arise through several distinct but interacting pathways, including autonomous misalignment, harmful human use, organizational failure, and competitive deployment. The severity of these pathways depends on factors such as capability, autonomy, external access, persistence, institutional safeguards, and the preservation of recovery capacity. The analysis is deliberately non-operational: it identifies causal conditions, empirically tractable intermediate quantities, and defensive research questions rather than procedures for causing harm.
Read the full abstract ▾
Artificial Intelligence
An Empirical Study of Agent Skills' Downstream Utility
Yu Cheng, Dehai Zhao, Zhongxin Liu et al. (+3 more)
Agent Skills package procedural guidance and resources for reuse, but a relevant Skill does not necessarily improve task performance. Existing studies characterize Skill content and evaluate downstream performance, yet provide limited explanations of how utility depends on content, execution configuration, and multi-Skill organization. We conduct an empirical study on 87 SkillsBench tasks, defining downstream utility as the pass-rate difference from No-Skill on the same tasks under the same model–harness configuration. We compare the same Skills across nine configurations, then examine alternative published Skills and organizations of fixed Skill sets under three selected configurations. We retrieve marketplace candidates from a curated corpus of 37,596 Skills. LLM-assisted analysis of content, execution traces, and final artifacts, followed by author review, relates provided support to actual use and task outcomes. The same Skills help some configurations and hurt others on 36.78% of tasks, with trajectories showing that recommended procedures can become an execution burden. Relevance rankings overlook more useful candidates. Within the evaluated candidate sets, reranking by support for required operations raises first-choice pass rates by 4.35–5.80 percentage points across the three configurations. We derive 17 authoring practices linking executable procedures to recovery, preservation of task requirements, and checks on final artifacts. Stage Plan and Dependency DAG outperform use order alone, with DAG's additional benefits concentrated in tasks supplied with five or six Skills. These findings guide developers to assess usable operation support, allow procedure adaptation while preserving task requirements, and make artifact dependencies explicit when organizing Skills.
Read the full abstract ▾
All authors Yu Cheng, Dehai Zhao, Zhongxin Liu, Qing Huang, Zhenchang Xing, Xiaoxue Ren
Artificial Intelligence · Cross-listed from cs.CR
CredLeakBench: Evaluating Credential Leakage and Recovery in LLM Agents
Rafid Ahmed, Joseph Fioresi, Mubarak Shah et al. (+1 more)
Language model
agents are increasingly deployed to automate everyday digital chores from managing emails and social media to handling banking and bills allowing users to step away from supervision. However, this capability also exposes sensitive information to phishing. Safe execution requires distinguishing malicious requests from genuine ones without simply refusing to act. Despite its practical importance, this problem remains underexplored and it is unclear whether current
agents or existing defenses can achieve it. To study this problem, we first propose CredLeak-Bench, a comprehensive
benchmark designed to evaluate how effectively and securely
agents automate human workflows when confronted with phishing and identity verification. The
benchmark covers both user-directed authentication and autonomous inbox monitoring, where
agents are not explicitly instructed to log in. It systematically varies deceptive cues and pairs phishing scenarios with legitimate counterparts, enabling joint evaluation of information leakage and utility on genuine tasks. Within a sandboxed environment, leakage is measured through actual submissions of information rather than
agents' self-reported behavior. Our evaluation reveals that all tested models are vulnerable to leakage.
Agents also disclose sensitive information during autonomous inbox monitoring, demonstrating that phishing can induce disclosure without a user request to authenticate. Furthermore, most evaluated mitigations that reduce leakage also impair performance on genuine tasks, exposing a security utility trade off in existing defenses. These findings show why reducing leakage alone is insufficient: effective defenses must prevent unauthorized disclosure while preserving legitimate task completion. CredLeak-Bench provides a controlled framework for measuring both objectives and evaluating progress toward secure, useful
agents.
Read the full abstract ▾
All authors Rafid Ahmed, Joseph Fioresi, Mubarak Shah, Yuzhang Shang
Artificial Intelligence · Cross-listed from cs.CV
4D-HOF: Hand-Object Flow Matching for Feed-Forward 4D Interaction Reconstruction
Shiqi Li, Sean Cho, Yijie Li et al. (+3 more)
Existing methods for 4D hand-object reconstruction often rely on costly per-sequence optimization, while generative approaches typically synthesize interactions from random noise, which can lead to unstable interaction prediction. We introduce 4D-HOF, a feed-forward framework that reconstructs 4D hand-object interactions from coarse but informative estimates produced by
vision foundation models. Concretely, we learn a conditional flow matching model that transports foundation-model-derived hand-object states toward an interaction
manifold, allowing the model to correct errors in translation, rotation, and
alignment in a feed-forward manner. A key advantage of our generative formulation is that it naturally enables test-time guidance within the transport process. Rather than applying a separate post-hoc optimization after reconstruction, we directly steer the evolving generative states using physical interaction constraints and observed 2D evidence, allowing the reconstruction to be refined as part of the generative process itself. By training the generative model on diverse
datasets, 4D-HOF generalizes robustly to challenging in-the-wild scenarios. Experiments on out-of-domain
benchmarks show that 4D-HOF achieves state-of-the-art performance, producing more stable and accurate 4D hand-object reconstructions.
Read the full abstract ▾
All authors Shiqi Li, Sean Cho, Yijie Li, Fengzhi Guo, Bowen Wen, Cheng Zhang
Artificial Intelligence · Cross-listed from cs.CL
IdeaAnchor: Teaching LLMs to Turn Literature into Research Ideas
Ziyu Chen, Yilun Zhao, Jiashuo Sun et al. (+3 more)
Scientific research often begins by synthesizing ideas from a set of related papers to identify gaps and formulate new directions. However, training
language models to perform this form of literature-grounded ideation remains challenging, as existing approaches based on prompting or feedback lack structured supervision for how papers should be synthesized. We introduce IdeaAnchor, a paradigm for training LLMs to perform research ideation using structured specifications as privileged signals. Each IdeaAnchor instance encodes how each input paper should be synthesized into a successful idea, including their functional roles, relationships, and target synthesis criteria. We build this paradigm by mining instances from published papers, capturing how real ideas emerge from prior literature. We then train models via demonstration, self-
distillation, and reinforcement learning, and further enhance generation with
retrieval at inference time. Experiments show consistent improvements in ideation quality. Our analysis reveals a functional decomposition: anchor-based training strengthens creative synthesis,
retrieval enhances detail elaboration, and combining both yields the best performance.
Read the full abstract ▾
All authors Ziyu Chen, Yilun Zhao, Jiashuo Sun, Yiling Ma, Manasi Patwardhan, Arman Cohan
Artificial Intelligence · Cross-listed from cs.RO
DepthWorld: 3D World Model for Robot Manipulation
Jai Bardhan, Josef Sivic and Vladimir Petrik
World models offer a data-driven alternative to traditional simulators for robotics, with applications spanning policy evaluation, improvement, and planning. All of these uses depend on faithful 3D geometry, yet current
video-based world models are trained on RGB alone and produce rollouts that look correct frame-by-frame but do not compose into a consistent 3D world. Closing this gap requires progress on two fronts: large-scale 3D supervision for manipulation, and an architecture that can absorb it without disturbing strong pretrained
video priors. We introduce a calibration pipeline that combines learned stereo depth with a joint factor graph, pooling all episodes collected from the same physical
robot to recover its shared kinematic parameters alongside per-scene extrinsics. Applied to the DROID
dataset, this yields DROID-3D, a calibrated 3D
dataset providing dense metric depth and recalibrated multi-view extrinsics (achieving <0.7 px reprojection error on 90% of episodes for external cameras). We then train DepthWorld, a Stable
Video Diffusion-based world model that jointly predicts multi-view RGB and depth via spatial latent tiling, leaving the pretrained Variational Autoencoder (VAE) unchanged. Depth supervision improves RGB prediction itself by +1.48 dB PSNR over an identical RGB-only baseline at equal training budget, while simultaneously yielding accurate metric depth for downstream
geometric reasoning.
Read the full abstract ▾
Artificial Intelligence
Sherpa: Teaching LLMs to Teach Adaptively
Weixian Xu, Yanzhe Zhang, Zora Zhiruo Wang et al. (+2 more)
Large
language models (LLMs) have become increasingly capable problem solvers, but being able to solve a problem is not the same as being able to teach it. Existing approaches to training LLMs as teachers rely on demonstrations, preference data, or predefined pedagogical criteria that specify what good teaching looks like. However, these signals are often not grounded in individual student learning outcomes, where effective teaching strategies can vary substantially across learners. To address this, we introduce Sherpa, a multi-turn reinforcement learning framework that instantiates multiple student archetypes with LLMs conditioned on distinct learning preferences and trains a teacher model to adapt its instruction by directly maximizing their learning outcomes. Teacher LLMs trained with Sherpa improve instructed students' performance across all archetypes by an average of 20.5 percentage points. Under MathTutorBench's evaluation, Sherpa raises the overall pedagogy score from 52.5% to 79.2%, indicating better teaching responses. Our human studies show that the trained teacher is preferred over the base model in 79.6% of pairwise comparisons. Together, Sherpa trains LLM teachers to adapt to diverse simulated students and become better aligned with human teachers, paving the road towards AI tutors teaching real students.
Read the full abstract ▾
All authors Weixian Xu, Yanzhe Zhang, Zora Zhiruo Wang, Changyu Chen, Diyi Yang
Artificial Intelligence
Agent in a Bottle: Can LLM Agents Turn Their Capabilities Into Cheap, Scalable Artifacts?
Ankit Sonthalia, Haritz Puerto, Alexander Rubinstein et al. (+2 more)
Large
language models (LLMs) can solve many narrow tasks, but querying them separately for millions of related instances can be prohibitively expensive. Can LLM
agents autonomously create cheaper solutions for such workloads? We call this ability "bottling": the ability to turn general capabilities into task-specific solutions that balance answer quality and amortised cost. We introduce BOTTLED, a
benchmark in which
agents receive an entire unlabelled workload and must complete it under fixed time, compute and LLM API budgets.
Agents choose their own approach, such as training a small model or writing a reusable program. Across ten models and three tasks, we find that strong zero-shot task performance does not reliably translate into strong bottling capabilities. Models with similar zero-shot scores can differ substantially after bottling, and 48 of 60 bottling runs score below the lower bound of the 95% confidence interval of their model's zero-shot performance. Moreover, 31 of 60 runs underperform the stronger of two small-model
distillation baselines with the same
token budget. Nevertheless, bottling can yield substantial savings: on query-product relevance classification, Opus 5 retains about 82% of its zero-shot macro-F1 at roughly 657 times lower reported cost. Bottling is also competitive with Jev, a "system one" model built especially for cheap, repetitive inference: Opus 5 on the same task recovers about 94% of Jev's macro-F1 at a quarter of Jev's projected full-workload cost. BOTTLED provides a basis for evaluating and improving
agents' ability to invest limited resources in reusable solutions for large, repetitive workloads.
Read the full abstract ▾
All authors Ankit Sonthalia, Haritz Puerto, Alexander Rubinstein, Martin Gubri, Seong Joon Oh
Artificial Intelligence · Cross-listed from cs.CL
AdvSim2Real : Training Web Agents Against Adaptive Prompt Injection in a Web World Model
Sarim Hashmi, Mukul Ranjan, Kshitij Mishra et al. (+3 more)
Web
agents complete user requests by reading and acting on pages that third parties write, so an instruction planted on a page can redirect the
agent away from the user's goal. The
agent cannot simply ignore the page, because the page also holds the values and controls the task requires. Current defenses fine-tune the
agent on injections fixed before training, and attackers that adapt to the trained model bypass them.
Adversarial training lets the attacker adapt but keeps the tasks fixed, so a task stops teaching once the
agent solves it. We introduce AdvSim2Real, which co-evolves a task curriculum, an injection adversary, and the
agent inside a frozen web world model. The curriculum is rewarded for tasks the
agent solves about half of the time, and the adversary only for a success flip, an injection that turns a judged success into a failure. Training in the simulator makes a 4B
agent both more capable and more robust: its completion rises with and without attacks, holds against a frontier-model adversary it never trained against, and its capability gain carries over to a real browser. On 150 web tasks, AdvSim2Real raises completion under this unseen adversary by 33.6% relative to the base
agent.
Read the full abstract ▾
All authors Sarim Hashmi, Mukul Ranjan, Kshitij Mishra, Mikhail Kuznetsov, Praneeth Vepakomma, Nils Lukas
Artificial Intelligence
VeriFine: Scaling Verification for Self-Improvement in Embodied Reasoning
Zewei Zhou, Rachel Luo, Yulong Cao et al. (+10 more)
Self-improving policies continually expose new failure patterns, changing what their judges must be able to verify. However, current fixed judges constrain both optimization feedback and the discovery of useful training examples, limiting further self-improvement. This challenge is even more acute in embodied
reasoning, where reliable evaluation must account for spatial grounding, causal
reasoning, and safety-aware decision-making. We introduce VeriFine, an
agent harness framework that scales verification through the co-
evolution of the policy, training curriculum, and judge. The Policy Improvement Loop uses a rubric judge to diagnose recurring failures, construct an adaptive curriculum, and optimize the policy. When progress plateaus and verification becomes a bottleneck, the Judge Improvement Loop selectively queries human guidance on informative failure cases and refines the judge through coactive calibration, in which humans and
agents resolve disagreements and converge toward the objective rubric of physical
reasoning. The revised judge then guides the next stage of data selection and policy optimization. Experiments on driving and
robot navigation tasks demonstrate continuous self-improvement in both policy and judge capability across reinforcement and supervised fine-tuning. These results show how scaling verification supports continuous self-improvement as policy failure patterns evolve.
Read the full abstract ▾
All authors Zewei Zhou, Rachel Luo, Yulong Cao, Chaowei Xiao, Chensheng Peng, Boyi Li, Thomas Tian, Zheng Lian, Yan Wang, Jiaqi Ma, Boris Ivanovic, Marco Pavone, Wenhao Ding
Artificial Intelligence · Cross-listed from cs.SD
WorldSonus: Bringing Sound to Worlds
Pengjun Fang, Jingyi Fa, Kam Man Wu et al. (+9 more)
Recent advances in world models have enabled increasingly realistic visual synthesis. However, these generated environments remain largely silent. Bringing sound to world models poses three core challenges: real-time generation to keep pace with interactive
video streams, interactive control to respond to mid-stream sound instructions, and spatially aligned stereo to reflect scene geometry and camera motion. To address these demands, we introduce WorldSonus, an interactive
video-to-audio framework designed for real-time spatial sound synthesis in world models. For real-time generation, WorldSonus employs a streaming causal autoregressive
diffusion architecture that synthesizes audio chunks at a low real-time factor (RTF) of 0.41. For interactive control, we incorporate an audio-centric captioning pipeline with chunk-indexed
prompt scheduling, enabling dynamic manipulation of sound events during generation. For spatial
alignment, we leverage high-quality stereo supervision curated from diverse stereo and ambisonic data. Extensive experiments demonstrate that while tailored for world models, WorldSonus generalizes effectively to open-domain
video-to-audio
benchmarks, matching or outperforming state-of-the-art bidirectional models in both acoustic quality and spatial
alignment. Project page: https://noizai.github.io/WorldSonus/
Read the full abstract ▾
All authors Pengjun Fang, Jingyi Fa, Kam Man Wu, Jiaming Wang, Haoyuan Huang, Yaguang Wu, Xiangjun Huang, Ziyang Ma, Weijia Chen, Hongyu Liu, Zeyue Tian, Qifeng Chen
Artificial Intelligence · Cross-listed from cs.LG
Reinforcement Learning with Conformal Action Sets: An Application to Sequential Recommendation
Wenwen Si and Honghao Wei
Sequential recommenders typically use a fixed slate size even though the number of useful alternatives changes within a session. We propose Reinforcement Learning with Calibrated
Pruning (RLCP), which adapts the retained action set using critic scores and an online threshold. The threshold is updated from binary feedback indicating whether the set contains an action in a proxy target. We prove a deterministic bound on the observed proxy miss rate along adaptive trajectories. To quantify the effect of
pruning on reward, we derive an exact decomposition of value loss into filtering and selection losses. Under explicit proxy and critic approximation conditions, this decomposition yields a finite session reward bound that also accounts for imperfect selection and set truncation, without requiring the learning parameters to converge. Experiments on KuaiRand-Pure and MovieLens 1M compare two RLCP implementations with four RL baselines. In each of the 19 configurations, at least one RLCP variant achieves the highest catalog diversity, reaching $1.11\times$ to $5.21\times$ that of the strongest baseline, with competitive session depth and no larger retained sets.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.RO
EgoLAP: Learning from Egocentric Human Data through Language-Action Reasoning
Lihan Zha, Shresth Grover, Tenny Yin et al. (+7 more)
Egocentric human data offer a path to scaling
robot learning beyond costly
robot demonstrations, yet the embodiment gap makes raw human trajectories a poor supervisory target for control. Our key insight is that, although low-level actions are embodiment-specific, their underlying motion intent can capture task-relevant structure that transfers across humans and
robots. We introduce EgoLAP, a VLA pre-training framework that jointly learns from human and
robot trajectories through a shared
language-based action chain-of-thought. EgoLAP expresses motion intent as structured, temporally abstracted
language actions and pairs them with motion-level
reasoning grounded in scene geometry, physics, and object affordances. Across extensive real-world and simulated experiments, EgoLAP transfers human experience to
robot control more effectively than alternative action representations and reaches 80.1% mean real-world task progress, a 2.3x performance gain over alternative action representations. Motion-level
reasoning also outperforms a composite
reasoning format that combines subtask, object-box, and visual-trace
reasoning.
Read the full abstract ▾
All authors Lihan Zha, Shresth Grover, Tenny Yin, Samuel M. Bateman, Hengkai Pan, Mengchao Zhang, Aykut Onol, Allen Z. Ren, Dhruv Shah, Anirudha Majumdar
Artificial Intelligence
Does an Agent's History Tell You When Compaction Will Hurt? A Modest, Bounded Effect on the TRACE Paired-Replay Corpus
Egor Pakhomov and Erik Nijkamp
Many long-horizon
agents compact their context on a global rule, usually a
token budget, blind to what the
agent was doing. We ask whether the
agent's recent behaviour predicts when a compaction will hurt. TRACE's public corpus of 590 harness-triggered AppWorld compaction boundaries replays each boundary from a re-executed prefix state under the pre-compaction context and under the summary, and records the burden of the next actions: calls that error or repeat a call already made. We find that pre-boundary history predicts post-compaction harm only weakly. An internally prespecified contrast by prefix placement is a wide null, and the naive "has-written" label behind it turns out to measure trajectory phase. The best extension-protocol trigger reaches held-out AUROC 0.66 (0.64 on the replicate's own label) against a same-boundary replicate of 0.72; the best frozen, interpretable trigger avoids 21% of harmful (positive-burden) boundaries while keeping 84% of compaction opportunities, and exceeds the random-rule expectation on count but not on burden mass (a post hoc comparison). Whether the best trigger beats a
token-budget rule at matched retention cannot be evaluated on the release. We state what corpora should ship to answer it.
Read the full abstract ▾
Artificial Intelligence
WorldSolver: Can LLM Agents Simulate the Physical Dynamics via Solver Generation?
Siru Jiang, Yongzhe Lyu, Shuo Lu et al. (+6 more)
LLM-based
agents are increasingly advancing scientific and engineering problem solving, with physics simulation emerging as a challenging yet practical testbed for reproducing complex physical phenomena with application in embodied AI, games and films. As the workhorse of such simulation, a solver computes how the state of a dynamic system evolves over time. Building such solvers requires physical understanding to identify appropriate models, mathematical
reasoning to formulate the underlying dynamics, and software engineering to implement them as executable code, yet this capability of LLM
agents remains underexplored. To this end, we introduce WorldSolver, a
benchmark of 168 simulation tasks derived from physical phenomena in 61 classic computer graphics papers, spanning 7 physical domains. Each task contains a code scaffold that provides a fixed simulation environment for the scene, with the solver implementation left for the
agent to complete. Specifically, we evaluate them along three dimensions: Execution Checks for successful execution, Visual Fidelity for reproducing the intended dynamic behavior in the rendered simulation, and Physical Plausibility for physics-grounded verification of the generated dynamics. Experiments on frontier
agents reveal that producing executable solvers is difficult itself, and satisfying visual and physical correctness is even harder. GPT-5.6-Sol and Claude-Opus-5 perform comparatively better than the other evaluated
agents, yet achieve overall scores of only 48.7% and 46.7%, respectively. WorldSolver is an early step toward agentic solver generation, and we hope it helps drive progress toward
agents that can faithfully simulate the dynamic physical world. Code is available at https://github.com/sirujiang/WorldSolver.
Read the full abstract ▾
All authors Siru Jiang, Yongzhe Lyu, Shuo Lu, Yubin Wang, Yuxiang Zhang, Yue Liao, Bin Wang, Jian Liang, Tieniu Tan
Artificial Intelligence
nanoMuse: An Open-Source Personal Agent for Every Device You Own
Guangyi Liu, Yong Liu and Jiangning Zhang
Assistants from 2011 answered and waited, and
agents from 2023 did a task and stopped. In September 2026 Meta's Muse showed an
agent for one person, with accounts, devices,
memory and a conversation that lasts, closed, in a vendor's cloud, in one country. Such an
agent is expected to act on a person's accounts and devices, remember them across weeks, speak first when it is worth it, and answer for what it did. It is a kind of software, not a model, and until now had no open counterpart. This report defines the personal
agent in five questions and three horizons. It reads how Muse is built from Meta's public record and a copy of its production
prompt, each statement marked by its source. It then presents nanoMuse, the open-source counterpart under the GPL-3.0, one
agent on every device a person owns, with hands on the phone's screen and the computer's. They share one conversation over a relay anyone can run; every action goes through a Sentinel,
memory is files the person can read, and the model is their choice. Its size and cost are given as estimates. What is open,
memory with provenance, an evaluation suite for the hands and an open model for them, is set out as a roadmap.
Read the full abstract ▾
Artificial Intelligence
ScienceClaw: Benchmarking Continual Self-Evolution of AI-for-Science Agents Across the Natural and Social Sciences
Mingda Zhang, Wenjin Liu, Tiesunlong Shen et al. (+6 more)
Large
language model
agents are accelerating scientific automation, yet verified executions rarely become persistent program-level improvements, and existing evaluations do not examine this process across sequential tasks in both the natural and social sciences. We formalize ScienceClaw as fixed-parameter program self-
evolution that unifies task solving, scientific verification, and program updates. ScienceClaw-Eval spans 23 disciplines and measures scientific correctness, evolutionary gain, retention, cross-
dataset transfer, and
evolution cost through sequential streams and independent reset evaluation. Our framework repairs executable workflows through multi-turn interaction, converts re-execution-verified failure–success trajectories into linked Skill and Operator candidates, and retains an update only when source-task replay reproduces the repair and independent scientific tasks improve. Code is available at https://github.com/beita6969/ScienceClaw.
Read the full abstract ▾
All authors Mingda Zhang, Wenjin Liu, Tiesunlong Shen, Zikai Xiao, Zhenghong Lin, Qing Xu, Erik Cambria, Xiaoying Tang, Haoran Luo
Artificial Intelligence
Coupled but Late: Turn-Taking Between Full-Duplex Speech Models in Unscripted Dialogue
Lichen Zhu, Yueqian Lin, Yiheng Wang et al. (+2 more)
Full-duplex
speech models are trained to converse with a person, but they are increasingly made to converse with each other, in self-play data generation,
agent societies, and model-based evaluation. In that loop no human absorbs a timing error: each model's turn-taking is the other's input. We ask what timing the loop settles into. Two PersonaPlex-7B instances exchange audio
tokens on a shared clock in unscripted conversation, and one floor-transfer rule is applied to them and to Switchboard. Their timing is coupled: re-pairing speakers across conversations destroys it. But the floor changes hands late, at a median of 400-560 ms against 137 ms for humans, and the last 120 ms of the partner's turn, where human projection places a tenth of its transfers, holds 1% of theirs. Delaying one direction of the channel shifts the response one-for-one and leaves the run-up to it empty, consistent with a reactive wait after the perceived end rather than the turn-end projection human timing requires.
Read the full abstract ▾
All authors Lichen Zhu, Yueqian Lin, Yiheng Wang, Hai "Helen" Li, Yiran Chen
Artificial Intelligence · Cross-listed from cs.CR
Secure Speculative Decoding for Large Language Models
Yichi Zhang, Zhiqi Wang, Neil Gong et al. (+1 more)
Speculative decoding accelerates inference for a large
language model (LLM), referred to as the target model, by first using a smaller model, referred to as the draft model, to generate candidate
tokens and then verifying them with the target model for acceptance or rejection. Prior studies primarily focused on the efficiency-utility trade-off of speculative decoding, e.g., lossy speculative decoding, leaving its security implications largely unexplored. In this work, we bridge this gap by providing the first systematic study of the security implications of speculative decoding. Through a large-scale measurement study, we reveal a pronounced security-utility asymmetry: across a wide range of lossy speculative decoding methods, improvements in inference efficiency come at a disproportionately high cost to security, with attack success rates for jailbreak and
prompt injection attacks increasing much faster than utility degrades. We then propose SecureSD, a new theory-guided speculative decoding method that enhances security while maintaining efficiency and utility. Specifically, our theoretical analysis reveals that security degradation primarily originates from the early
tokens generated by the draft model. Motivated by this insight, SecureSD applies a stricter verification criterion to draft-model
tokens at early decoding positions. Extensive experiments on both security and utility
benchmarks demonstrate that SecureSD significantly improves security while preserving efficiency and utility compared to existing speculative decoding methods.
Read the full abstract ▾
All authors Yichi Zhang, Zhiqi Wang, Neil Gong, Yuchen Yang
Artificial Intelligence · Cross-listed from cs.LG
Principled Under Pressure: Post-Training Decides Whether LLMs Act on Their Own Moral Judgment
Orion Reblitz-Richardson
Language models increasingly act as
agents. An
agent that says an action is wrong and then takes it anyway is a different failure from one that does not know better, and evaluations of stated values cannot see it. We build a pre-registered panel of 248 scenarios across five kinds of pressure. Each scenario is posed twice to the same model, once as the
agent choosing what to do and once in the third person asking which option is right, so the model's own judgment is the reference. Every scenario has a twin with the pressure removed, and every model gets a positive control in which its operator orders the violating action, so that a missing gap can be told apart from a blind instrument. On OLMo-3-7B-Instruct, the model takes the action it judged wrong on about one in five pressuring scenarios, more often than on the same scenarios with the pressure removed. Across four instruct models the gap depends on the post-training recipe: OLMo-3 and Meta's Llama-3.1-8B-Instruct carry it; Tulu 3 shows none on the whole panel (above about 0.01 in probability) or on its own most-pressuring scenarios; Qwen2.5-7B-Instruct shows none on the whole panel (above about 0.02) and is unresolved on its own (0.083, -0.028 to 0.195). Meta's recipe and Ai2's Tulu 3 start from the same Llama-3.1 weights, and only Meta's carries the gap. Reading a chat model outside its chat template reverses the sign of its gap with nothing at stake (-0.038 against +0.055 under the template on OLMo-3), a distortion present on two of three recipes. On both models that carry it,
reasoning about the stakes before acting moves the choice back toward the model's own judgment, against a same-length non-moral task, with or without the pressure; on OLMo-3, naming the norm at stake does about a third of that. The gap is a measurable target for post-training recipes, not a fixed property of pretrained weights.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.LG
MemFLoRA: Memory-Floor LoRA for CNN Adaptation at the Edge
Mehmet Emre Akbulut, Johannes Geier and Ulf Schlichtmann
On-device learning is necessary when the model encounters user-,sensor-, or environment-specific shifts after deployment. Although parameter-efficient fine-tuning (PEFT) methods, particularly Low-Rank Adaptation (LoRA) variants, enable efficient adaptation at the edge, the limiting resource for Convolutional Neural Network (CNN) adaptation is often not the number of trainable parameters but the activation state that must be retained until the backward pass. This paper introduces
Memory-Floor LoRA (MemFLoRA), a low-rank CNN adapter built around a
memory-first design principle rather than a direct application of
transformer-oriented LoRA. Instead of merely reducing trainable weights, we define an activation-
memory-floor criterion: trainable backward computations must not depend on full-width layer inputs. The resulting adapter freezes the down-projection, trains a scale-matched up-projection, and combines eval-mode backbone normalization with activation-minimal backward rules, reducing saved state to the low-rank branch. Evaluated on three Human Activity Recognition (HAR)
datasets and two CNN backbones under subject, body-location, and sensor-placement shifts, MemFLoRA reduces saved-activation
memory by 98.5-98.7% and peak training-state
memory by 94.9-97.3% relative to full fine-tuning, while matching or exceeding CNN PEFT baselines.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.CR
Semantic Behavioral Watermarking: Paraphrase-Robust and Forgery-Resistant Provenance for LLM Agents
Suxin Ji, Hungtao Wan, Shaoxuan Chen et al. (+1 more)
Behavioral watermarking embeds an owner identifier in an LLM
agent's high-level action choices, giving provenance without touching output
tokens. Prior
agent watermarks break in two ways. First, all three prior schemes bind the watermark to the exact action symbol, so renaming a tool desynchronizes decoding even when the observation is untouched; in AgentMark's own robustness test, paraphrasing the observation alone drops bit-recovery to 16.8%. Second, every prior
agent watermark studies only removal: none asks whether an adversary can forge a trajectory that verifies as someone else's, a question answered affirmatively for text watermarks (Jovanović et al., 2024). We present Semantic Behavioral Watermarking (SBW): watermarking over semantic action clusters under history conditioning, with the public-cluster bin replaced by keyed collision-resistant binning whose fresh-bucket assignment is provably unpredictable in the random-oracle model. Across five
agent models (3B-14B, four vendors) and three encoders the ordering holds on both
benchmarks: on ToolBench (600 trajectories per model) detection under rewriting is 0.49-0.66 for cluster-level versus 0.05-0.17 for exact-symbol at a permutation-calibrated 1% FPR, at 72-83% choice agreement against 22-27% for logit biasing; on ALFWorld (100 episodes per model) it is 0.92-0.97 versus 0.00-0.01. Keyed binning takes adaptive forgery from 100% to the false-positive floor at the primary operating point (bge, r=64). We also mark the boundary that guarantee does not cover: when the adversary copies the victim's own steps, shuffled splicing is neutralized (0.000 on Qwen2.5-3B) but chained replay remains at 0.76-0.98 across the five models, reported as open. Paraphrase robustness costs about half of the per-step watermark capacity. Code is available at https://anonymous.4open.science/r/SBW-
Agent-Watermark.
Read the full abstract ▾
All authors Suxin Ji, Hungtao Wan, Shaoxuan Chen, An Zhang
Artificial Intelligence
ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding
Hanjun Luo, Xiucheng Zhang, Zhuoning Xu et al. (+4 more)
As coding
agents increasingly undertake real-world work autonomously, judging whether their risk treatments are warranted has become important. Existing work evaluates related
agent behaviors from separate perspectives, but lacks a systematic framework for unifying these behaviors. To bridge this gap, we introduce ParanoiaEval, the first
benchmark for unified evaluation of risk-treatment capabilities in coding
agents. Grounded in the well-established Avoidance-Transfer-Mitigation-Acceptance framework in software engineering risk management, ParanoiaEval operationalizes its 4 fundamental treatments for coding-
agent settings and contains 200 evidence-controlled repository-level task pairs, each differing only in treatment-defining evidence. We further introduce dedicated metrics for risk-treatment violations and evidence responsiveness, using a human-calibrated agentic judge for reliable evaluation. Large-scale experiments on 8 representative models and a post-hoc human study reveal that (I) unnecessary risk treatment occurs in 11.2%-58.7% of runs despite explicit evidence, with substantial variation across
agent configurations; (II) stronger task capability does not ensure more appropriate risk treatment, while treatment violations substantially harm developers' experience, establishing risk treatment as an independent capability dimension; and (III)
agents exhibit systematic patterns consistent with established risk-management findings, suggesting that knowledge from human practice can guide the diagnosis and improvement of this capability.
Read the full abstract ▾
All authors Hanjun Luo, Xiucheng Zhang, Zhuoning Xu, Zhimu Huang, Yingbin Jin, Xinfeng Li, Hanan Salam
Artificial Intelligence · Cross-listed from cs.CV
Selective Transfer of RL Updates for Visual Reasoning
Suxin Ji, Hungtao Wan, Mingjun Liu et al. (+1 more)
Model merging provides a training-free way to transfer
reasoning capabilities from
language models to
vision-
language models (VLMs), but endpoint-based transfer can conflate pre-existing model differences with changes acquired during
reasoning post-training. We instead formulate capability transfer around the training-stage update, isolating the parameter changes induced by reinforcement learning (RL). Yet transferring this update in full remains suboptimal: we find that its components differ substantially in cross-model transferability, with dominant directions transferring more effectively than the complete update. Based on this finding, we introduce Selective-RL, which isolates the RL-stage update, retains its dominant
matrix-wise directions with magnitude preservation, and transfers them to the
language modules of a VLM. Across three model families and five visual-
reasoning benchmarks, Selective-RL improves full-update interpolation in 12 of 15 comparisons, including an 8.55 percentage-point MathVision gain on the Qwen recipient. Matched controls show that update magnitude or arbitrary low rank alone does not reproduce these gains. These results highlight a distinction between what is acquired during post-training and what remains transferable across models, providing a training-stage perspective on cross-model capability transfer. Code is available at https://anonymous.4open.science/r/selective-rl.
Read the full abstract ▾
All authors Suxin Ji, Hungtao Wan, Mingjun Liu, An Zhang
Artificial Intelligence · Cross-listed from cs.SE
A Case Study in Assuring AI-Written Software
Lindsey Ferris and Sierra Bonilla
Software-engineering
agents can enable people without formal software training to build systems they could not otherwise implement and simultaneously can produce more code than even experts can meaningfully inspect. In both cases, exhaustive code review is not reliable as the sole basis for human control. We report a case study of a production healthcare platform built through coding
agents and governed by an operator without formal software-engineering training. Over time, its workflow grew into a human-led meta-
agent system where one
agent wrote code, other
agents supervised and reviewed it, and project rules carried lessons forward. The operator found that tests, monitors and reviewing
agents used to supervise the system were fallible. Some monitors measured proxies rather than outcomes, some audits failed silently, missing checks disappeared from reported results and one automated repair caused operational disruption. In this case, human control depended on keeping the intended outcome, the evidence used to judge it, the
agents' permissions and the final human decision were all tied to the same underlying objective.
Read the full abstract ▾
Artificial Intelligence
SquidAgent: Parallelize Wisely, Coordinate Efficiently
Yexiong Lin, Shanshan Ye, Yu Yao et al. (+3 more)
LLM-based
agents solve complex multi-step tasks, but sequential execution incurs substantial latency. In principle, parallelizing work across multiple
agents should yield near-linear speedups. Yet existing parallel multi-
agent systems often run slower than a single-
agent baseline. We attribute this gap to two hidden costs that parallel execution incurs but a serial
agent avoids. First, there is a re-exploration cost: redundant effort spent by parallel workers reconstructing context that the orchestrator already possesses, such as prior decisions, that would otherwise be inherited implicitly in a serial execution. Second, there is an
alignment cost: the overhead required to reconcile inconsistencies across independently generated outputs. We thus derive a principled decision criterion: a layer should be parallelized only when its critical-path cost, plus re-exploration and
alignment overheads, is lower than the corresponding serial cost. While this criterion is naturally expressed in wall-clock time, we observe that LLMs are poorly calibrated when asked to estimate task duration. To address this, we instead measure cost in predicted output
tokens, which we empirically find LLMs can estimate substantially more reliably than wall-clock time. Building on this
token-based criterion, we propose SquidAgent. It estimates all
token budgets in a single planning step, forks each worker directly from the orchestrator's session to eliminate re-exploration cost, and replaces post-hoc reconciliation with a pre-generated shared convention block that converts
alignment into a bounded upfront cost. A deterministic scheduler then applies the criterion layer by layer. Empirically, SquidAgent achieves a 2.2$\times$ mean throughput improvement and a 2.6$\times$ mean wall-time speedup over Claude Code, and a 2.0$\times$ throughput improvement over the strongest multi-
agent baseline.
Read the full abstract ▾
All authors Yexiong Lin, Shanshan Ye, Yu Yao, Zhen Fang, Bo Han, Tongliang Liu
Artificial Intelligence · Cross-listed from cs.RO
HygieneRoboBench: Benchmarking Hygiene-Aware Planning for Household Robots
Yurun Chen, Josh Qixuan Sun, Jason Qin et al. (+4 more)
Contact with contaminated objects can spread hazards through a household
robot's grippers, tools, and shared surfaces, while new contacts can make an existing plan unsafe. Existing
benchmarks do not jointly assess how planners identify hygiene risks from contact history and plan safe continuations after new contact events. Planners must do so within time and resource limits while respecting user priorities. We introduce HygieneRoboBench, with 624 instances across 134 task families, to evaluate safe resolution of household tasks from a given execution history. Tasks capture contamination through two grippers and shared objects, treatment costs, and user priorities. We combine controlled history, profile, and event comparisons with independent plan evaluation. These assess safe resolution, cost efficiency under user priorities, and responses to contact events. Evaluation of LLM-based and symbolic planners shows that safely completing a task does not guarantee the lowest execution costs under the user's priorities. To address this problem, we introduce Hygiene-NSP. It combines LLM-based grounding, contact-history reconstruction, and CP-SAT to jointly plan hygiene treatment and task execution under user priorities. Hygiene-NSP achieves safe resolution and optimal safe resolution rates of 94.4% and 90.4%, respectively. Both rates are higher than those of the evaluated baseline planners on the full
dataset. Project page: https://euron-zc.github.io/HygieneRoboBench/.
Read the full abstract ▾
All authors Yurun Chen, Josh Qixuan Sun, Jason Qin, Chengtai Li, Tianyi Wang, Mark Crowley, Wentao Zhu