Artificial Intelligence · Cross-listed from cs.LG
RLTLDR: Self-improvement by Internalizing Self-generated Feedback
Michael Kirchhof, Eleonora Gualdoni, Andrew Szot et al. (+6 more)
The common paradigm of reinforcement learning with verifiable rewards (RLVR) is to let
agents make multiple attempts at a task, and optimize towards the successful ones. This becomes problematic in the realms of self-improvement, where tasks are so difficult that the
agent has a low or even no chance of success, and where there are no teacher models or example solutions to distill from. In this paper, we introduce RLTLDR. After each failed attempt, we show the policy the verifier outputs and let it write its own feedback, in the form of a single TL;DR insight. The next rollout is conditioned on all previous insights, and we sequentially sample rollouts until a solution is found. Moreover, we enable backpropagation on the in-context insights to internalize a direct task to insight mapping. On challenging tool-calling and coding
datasets (filtered to Pass@128=0), standard GRPO training of a Qwen 3.5 9B Thinking policy stays flat at a Pass@1 of 0% to 1%. RLTLDR breaks through this learning barrier, achieving a Pass@1 of 14-31% with insights in context during training and, crucially, 12-13% when no insight is in context at eval time. We identify that the key is the task to insight internalization. To study this further, we reduce our approach to SFTLDR, training only on (task, insight) tuples, without showing or backpropagating on any rollouts. Training on only 4k of these tuples recovers almost the full performance of RLTLDR and classical SFT on full rollouts. This demonstrates a promising compacted training paradigm of the form "on this sort of task, keep this sort of thing in mind", which we hope to inspire future research on.
Read the full abstract ▾
All authors Michael Kirchhof, Eleonora Gualdoni, Andrew Szot, Khashayar Gatmiry, Aryo Lotfi, Abbas Kazerouni, Omar Attia, Sanjoy Chowdhury, Alexander Toshev
Artificial Intelligence · Cross-listed from cs.LG
ProCTI: Prototype-Refined Global Conditioning for Diffusion-Based Time Series Imputation
Fariza Rashid, Duc Van Le, Rahat Masood et al. (+3 more)
Time series imputation has progressed from statistical and deep learning approaches to
diffusion-based models, which have shown strong recent performance. Existing
diffusion-based methods typically condition the reverse process using local contextual information from the current or neighbouring windows. Meanwhile, global
dataset-level structure often remains implicit, limiting performance when local observations are sparse, noisy, or unrepresentative. To address this issue, we propose ProCTI, a
diffusion-imputation framework that augments local conditioning with retrieved global
dataset-level priors through learned prototypes. A hybrid conditioning mechanism integrates this global context with local signals during reverse
diffusion, enabling more accurate reconstruction under varying missingness scenarios. Experiments across multiple
benchmark datasets show that ProCTI outperforms strong baselines overall under random missingness, while remaining competitive under attribute-wise missingness. Furthermore, we use a latent-regime data model to characterise the precise conditions under which prototype-derived global conditioning provably improves imputation. We support this with a general theoretical analysis of local-global conditioning.
Read the full abstract ▾
All authors Fariza Rashid, Duc Van Le, Rahat Masood, Gustavo Batista, Aruna Seneviratne, Suranga Seneviratne
Artificial Intelligence · Cross-listed from cs.DC
SPLASH: Switching Parallel Layouts of Attention with Seamless Handoff for LLM Serving
Chuan Liu, Shuoming Zhang, Zhicheng Li et al. (+6 more)
No single way of parallelizing
attention serves large
language models well under all loads. Low concurrency favors tensor parallelism, many independent requests favor data-parallel
attention, and long
prompts favor context parallelism.
Reasoning, agentic, and RL-rollout workloads make a fixed choice untenable: a batch that begins as many short requests ends as a few very long ones, so the best layout changes while the same requests run. Serving engines nevertheless fix one layout at launch, because changing it has meant draining requests and restarting workers. We present SPLASH, a serving system that switches the parallel layout of
attention while requests are running. It builds on one observation: modern
attention, with few or no KV heads, decouples where a request's KV cache lives from how
attention weights are sharded. This has two consequences. First, layouts differ only in who owns the weights and the cache, and most of that state already sits where the next layout needs it; SPLASH reuses it, moves the rest in the background of ongoing inference, and hands off at a batch boundary, making a switch nearly free: its median overhead is under 0.51% of the step it runs in. Second, the decoupling exposes a layout that existing engines lack: Decoupled Ownership Parallelism (DOP) shards
attention weights as tensor parallelism does while keeping each request's cache on a single owner as data-parallel
attention does. DOP replicates neither, offers 27-60% more KV capacity than data-parallel
attention, and gives the scheduler a choice when KV
memory limits admission. A transition-aware scheduler follows the best of the four layouts as load changes. On B200 GPUs serving GLM-5.3, SPLASH improves end-to-end serving throughput by 1.3-1.73x over fixed-layout deployments, and the same layout regimes appear with DeepSeek-V3.2 on H200 and GLM-5.3-Flash on DCU.
Read the full abstract ▾
All authors Chuan Liu, Shuoming Zhang, Zhicheng Li, Qianqi Sun, Ruiyuan Xu, Qiuchu Yu, Xiyu Shi, Huimin Cui, Jiacheng Zhao
Artificial Intelligence · Cross-listed from cs.CL
Correct, Don't Delete: Mitigating Emergent Misalignment with Corrective Supervision
Jacob Epifano
Fine-tuning a
language model on a narrow set of harmful demonstrations, such as bad medical advice, can make it broadly misaligned on unrelated questions, a phenomenon known as emergent misalignment (EM). The usual defense is to find the offending rows and delete them, but a row locator failed our held-out test and deleting rows helps less than expected. We ask a different question: given a fixed set of poisoned rows, is it better to correct them than to remove them? We fine-tune Qwen2.5-14B-Instruct on a mixture of bad medical advice and benign chat data, select a quarter of the poison rows in advance, and either delete them or replace each with a corrected answer to the same
prompt, keeping everything else the same. Replacing the rows cuts the EM rate by about a third and improves answers on held-out medical questions, while deleting the same rows has little measurable effect. The advantage is larger when half the poison rows are corrected, and it holds on a second base model and a second misaligned model organism. The content of the replacement appears to matter: paraphrasing the rows while keeping their bad advice shows no clear benefit, and the correct answers distributed with the
dataset appear to do about as well as our rewriter's. Realigning an already-poisoned model with further fine-tuning is known to work, but which data does the work has not been compared directly. We find that a short round of training on corrections beats the same amount of training on generic chat data, that corrections on other medical
prompts do roughly as well as corrections of the poisoned
prompts themselves, and that instructing the correction writer to model a careful, harm-avoiding assistant adds no measurable benefit over plain corrections. In the settings we tested, correcting harmful training data reduces EM more than deleting it.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.SD
AS$^2$D: Accelerating On-Demand Audio Understanding on Mobile Devices
Yunzhe Li, Kyoungjun Park, Hongzi Zhu et al. (+1 more)
Speculative decoding accelerates autoregressive generation by using a smaller drafter to propose
tokens for batched verification by a larger target. However, conventional speculative decoding couples drafting to the target's evolving verified prefix, serializing drafting and verification. We ask whether this dependency is necessary for source-conditioned generation. Our key observation is that, for audio
language models, the input audio and user request can provide useful speculative candidates without following the target's evolving text prefix. We propose AS$^2$D (Audio Speculative Speculative Decoding), which enables target-decoupled drafting: an audio-conditioned drafter follows its own generation history while the target independently verifies and corrects ready candidates. Without usable candidates, the target advances alone. Thus, target feedback determines which candidates are committed but no longer determines when the drafter can make progress, enabling drafting and verification to proceed concurrently while retaining target-side verification and correction. We implement AS$^2$D in MNN for Android and evaluate two target models across four phones, seven
datasets, and three tasks covering 12.2 hours of audio. Across four phones, AS$^2$D improves pooled ASR throughput by 42-76% over target-only decoding, while only 5.7% of evaluation windows are slower than target-only, compared with 58.1-63.0% for speculative baselines. For ASR, AS$^2$D reaches 97.33-98.20% of a hindsight per-window oracle's pooled throughput over the evaluated drafter/budget catalog. Native on-demand execution with a 7B target achieves up to 78% higher throughput than target-only. These results show that source-conditioned audio generation can relax the conventional dependence of speculative drafting on the target's evolving output prefix, exposing substantial parallelism for efficient inference.
Read the full abstract ▾
All authors Yunzhe Li, Kyoungjun Park, Hongzi Zhu, Lili Qiu
Artificial Intelligence · Cross-listed from cs.SE
Independent Verification Paths Are Not Independent: A Case Study of Common-Mode Failure in a Satellite Catalogue Pipeline
Fabio Rovai
A common safeguard for a data pipeline is redundant computation: derive each published number by two routes built on different technology and refuse to exit when they disagree. We report one such gate failing, in a cross-catalogue integrity study of two open registers of Earth-orbiting objects. A gate comparing a set-based Python path with SPARQL queries over the emitted RDF graph printed ALL CROSS-CHECKS AGREE on seven counts. Three were wrong, one overstated more than fourfold (932 against 220). Both paths imported the same constants, which encoded a misreading of the source's status vocabulary, so the error was common-mode and the gate could not see it. We give the mechanism, an object-level ledger reconciling every figure, and three checks that go back to the source's documentation, measured on the defective code and on its correction. We then checked that correction against each object's phase history, held in a source file the pipeline never read. The correction was also wrong: 42 of its 261 disagreements are artefacts, and none of our three checks flagged them. Finally, in a controlled replication with three pinned models and tools disabled, 72 of 75 paths generated on request as independent checks computed the defective count, 29 of 30 even when the
prompt carried the source's own definitions of the codes. The evidence is one pipeline and one defect family. Within it, redundancy verified implementation, and the errors that reached publication were errors of meaning.
Read the full abstract ▾
Artificial Intelligence
XU-RS: Explaining Credal Width in Random-Set Language Models
David Achara, Maryam Sultana, Alexander D. Rast et al. (+1 more)
Uncertainty estimates tell us how unsure a model is, but not why. Without knowing which parts of an input influences a model's uncertainty, we cannot tell whether that uncertainty score depends on input features that are relevant for the task. We study this problem in randomset classifiers built using pretrained
language models. These classifiers assign probability to individual answers and to groups of answers, producing lower and upper probabilities for each answer; The difference between these probabilities, called credal width, is used to represent epistemic uncertainty about an answer arising from limited training data. We propose XU-RS, a framework that attributes an answer's credal width to the input
tokens (words or word pieces) supplied to a
language model. XU-RS uses Expected Gradients (a standard feature attribution method) to estimate how input
tokens contribute to credal width. The proposed framework is evaluated on a MedQA
dataset using SmolLM3-3B and Llama-2-7B models, demonstrating that setting the
embedding of a
token ranked highly by XU-RS to zero (zero-masking) causes larger changes in credal width than zero-masking randomly selected
tokens. In addition, we show that normalisation can cause other answer groups to influence an answer's width, reveal how
token attribution can mask numerical errors, and provide diagnostic checks to verify whether a
token ranked highly by XU-RS meaningfully explains model uncertainty.
Read the full abstract ▾
All authors David Achara, Maryam Sultana, Alexander D. Rast, Fabio Cuzzolin
Artificial Intelligence · Cross-listed from cs.RO
Credit-Guided Policy Improvement for Test-time Adaptive Vision-Language Navigation
Yang Li, Sijia Zhang, Yihan Li et al. (+4 more)
Test-time adaptation for
vision-
language navigation (TTA-VLN) enables pretrained policies to adapt online to unseen environments using only test-time observations and interaction history. However, distribution shifts can distort local action preferences and lead to off-course decisions. Existing methods rely on predictive uncertainty, trajectory-level feedback, or accumulated adaptation experience to correct such deviations. These signals, however, do not directly reveal whether an executed action supports instruction-guided progress toward the goal. Moreover, a plausible corrective signal does not guarantee a reliable policy update. The key challenge is thus twofold: identifying interactions that support goal-directed improvement and determining whether the resulting updates are worth retaining. We observe that each executed action induces an immediate observation transition, providing evidence of its local consequences. Based on this insight, we propose Credit-Guided Policy Improvement (CGPI), which recovers signed, reference-relative decision credit from action-induced observation transitions without external outcome feedback. With the pretrained
navigation policy frozen, CGPI uses this credit to propose lightweight adaptation updates and verifies them against prior credit-supported interactions. Updates are retained only when supported and rolled back otherwise. CGPI achieves consistent gains across the evaluated VLN
benchmarks and
navigation backbones, while qualitative
robot trials further illustrate the feasibility of zero-shot sim-to-real transfer.
Read the full abstract ▾
All authors Yang Li, Sijia Zhang, Yihan Li, Aming WU, Zihao Zhang, Ziju Han, Yahong Han
Artificial Intelligence
FOCUS: Training-Free Decision-Preserving Context Compression for LLM Agents
Shantanu Dixit, Anson Bastos, Xuchao Zhang et al. (+2 more)
LLM
agents accumulate interaction histories that grow linearly with task length, causing quadratic inference cost scaling and performance degradation from
attention dilution. Existing context-compression methods learn what to discard offline: by contrastively optimizing guidelines, distilling compressors, or training compression policies. This incurs a substantial cost. Further, the compression policy is learned a priori and is not dynamically conditioned on the evolving test-time trajectories. In this paper we ask a complementary question: Which past interactions causally shape the
agent's future decisions? We recast context compression as a causal decision preservation problem over discrete interaction units and introduce FOCUS, a training-free context compression framework that operates entirely at test time. Our method requires no offline data collection or fine-tuning, and is architecture-agnostic, attaching to any closed-API frontier model as a modular compression layer. We evaluate FOCUS on diverse agentic
benchmarks including API and tool-calling, QA, web domain and multi-turn dialogue. Our method establishes new state of the art performance, cutting peak context by up to 48% and dependency by 73% while improving task success by up to 8.9 percentage points over uncompressed execution.
Read the full abstract ▾
All authors Shantanu Dixit, Anson Bastos, Xuchao Zhang, Chetan Bansal, Saravan Rajmohan
Artificial Intelligence
Rational Clarification by Assistive Agents via Value-of-Information Reasoning
T. Duy Nguyen-Hien, Yee Whye Teh, Wee Sun Lee et al. (+1 more)
Users of
language-based assistive
agents often make ambiguous requests. In response, an assistant can either directly act on its interpretation of the request — risking misalignment with the user — or ask a clarifying question. Which option is the most safe and helpful? A common approach is to ask questions that minimize uncertainty about the user's intent until a threshold is reached. However, this neglects the impact of uncertainty reduction on downstream performance, the costs of asking versus acting immediately, and the possibility that users may provide corrections without being asked. To navigate these trade-offs, we introduce Rational Enquiry via Value-of-Information
Reasoning (REVOIR). REVOIR makes clarification decisions via inference-time
reasoning about the value-of-information of a question, which captures the expected improvement in task reward due to the answer received. In two assistive tasks — ambiguous question answering (CondAmbigQA) and preference-aligned household task planning (ADAPT) — we show that REVOIR achieves greater success with fewer questions than approaches based on prompting, chain-of-thought, fine-tuning, or information gain, improving preference satisfaction on ADAPT by 13-15% over a fine-tuned clarification policy while requiring no training and asking five times fewer questions. Furthermore, when the assistant can receive cheap user corrections after acting, REVOIR naturally infers that asking questions is not always efficient, demonstrating the adaptivity of our approach. In contrast, we find that vanilla
reasoning agents fail to adaptively clarify user requests, and request fewer clarifications as
reasoning effort increases.
Read the full abstract ▾
All authors T. Duy Nguyen-Hien, Yee Whye Teh, Wee Sun Lee, Tan Zhi-Xuan
Artificial Intelligence · Cross-listed from cs.LG
ReLMem: Learning Recurrent Memory for Longitudinal EHR Modeling
Zijie Meng, Xiwei Dai, Yingying Zhang et al. (+3 more)
Longitudinal electronic health record (EHR) modeling requires integrating new visits with an expanding patient history. Yet the continual accumulation of clinical information imposes increasing computational and
memory costs on large
language models (LLMs) when they process and retain complete patient histories. A practical alternative is visit-wise recurrent compression, which incorporates each incoming visit into a compact, continually updated patient
memory. However, under a fixed
memory budget, successive updates must integrate new information without progressively losing critical historical evidence needed to subsequent tasks. To address this challenge, we introduce Recurrent Longitudinal
Memory (ReLMem), a framework that learns to maintain fixed-capacity patient
memory for efficient downstream prediction with a frozen LLM. ReLMem equips this LLM with lightweight compression adapters to recurrently update the
memory from its previous state and each incoming visit, without rereading earlier records. Specifically, we develop a multi-granularity optimization strategy to preserve task-relevant information throughout recurrent updates and support downstream prediction from the final
memory. The intermediate supervision aligns
attention outputs from compressed
memory and the full history under identical queries, while prediction supervision minimizes cross-
entropy with ground truth answers conditioned on the final
memory. On EHR-based medication prediction, ReLMem approaches the F1 scores of full-history baseline while reducing average retained historical storage by 97.1%. Under the same
memory budget, it improves macro- and micro-F1 over the strongest compressed-
memory baseline by 4.66 and 4.75 percentage points, respectively. These results highlight the value of learning recurrent patient
memory for efficient longitudinal EHR modeling.
Read the full abstract ▾
All authors Zijie Meng, Xiwei Dai, Yingying Zhang, Jian Wu, Xian Wu, Zuozhu Liu
Artificial Intelligence · Cross-listed from cs.SD
Learning as Deepfakes Evolve: RF-Prompt for Continual Audio Deepfake Detection
Yuankun Xie, Xiaoxuan Guo, Xiaopeng Wang et al. (+3 more)
Continual audio deepfake detection requires learning newly emerging deepfake methods while retaining discrimination of previously encountered
speech. Existing
dataset-incremental evaluation changes both real-
speech domains and deepfake mechanisms, making their effects difficult to distinguish. We construct five task organizations over identical training, development, and evaluation pools to study these factors under a controlled sample budget. Our proposed Real-Anchored Mechanism-Incremental (RAMI) protocol reflects the practical setting in which available real
speech provides a recurring mixed-domain reference while new deepfake mechanisms arrive incrementally. We further propose RF-
Prompt, an asymmetric continual
prompt-learning method that preserves reusable real-
speech knowledge through a shared real
prompt and expands mechanism-specific knowledge through inherited fake experts with orthogonal residuals. Input-adaptive soft fusion combines the accumulated experts into a fixed number of injected
tokens without requiring task identity at inference. On RAMI, RF-
Prompt achieves 10.110% average EER and 10.370% pooled EER, outperforming all evaluated continual-learning baselines. Across the five controlled protocols, RAMI yields the lowest common-average and pooled EER. Component ablations, limited-data experiments, and cross-backbone evaluations further validate the proposed design.
Read the full abstract ▾
All authors Yuankun Xie, Xiaoxuan Guo, Xiaopeng Wang, Siqing Qin, Shaole Li, Kong Aik Lee
Artificial Intelligence · Cross-listed from cs.CV
FedSocket: Recipient-Executable Knowledge Exchange for Heterogeneous Multimodal Federated Learning
Xinyuan Zhao
Federated knowledge must remain usable by recipients with different modalities, private architectures, and tasks. We present FedSocket, which makes recipient execution a design requirement of the exchanged model. A shared Q combines recipient-computable inputs, task-owned outputs, and ownership-aware aggregation, connecting heterogeneous private models through a common prediction interface. Private models teach local Q copies; the returned Q supports local learning and Joint inference, with only Q parameters and counts exchanged. Across six
datasets, FedSocket improves missing-modality recipient accuracy over Local by 14.44 and 15.51 percentage points on MELD and UCF-51. Under matched inference capacity, Joint exceeds independent ensembles by 11.06 points in UCF-51 accuracy and 4.87 points in mean bidirectional Flickr30k R@1. Joint also improves over Q alone on all four heterogeneous endpoints, demonstrating the value of combining local and exchanged predictions. Teacher controls, sharing-path interventions, and component factorials identify the roles of supervision, sharing, and deployment. FedSocket makes exchanged knowledge directly usable from
federated training to recipient inference.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.CV
TReVS: Integrating Textual Relevance and Visual Saliency for Efficient Vision-Language Model Token Pruning
Jing Wang, Zhiping Wu, Dongdong Ren et al. (+3 more)
Vision-
Language Models (VLMs) excel at visual understanding and
reasoning but often incur substantial inference costs due to the large number of visual
tokens. Recent visual
token pruning methods increasingly follow a two-stage paradigm: they first remove visually redundant
tokens after the
vision encoder and then discard
tokens irrelevant to the textual query within the Large
Language Model (LLM). However, since the first stage typically relies solely on
vision-encoder saliency, it may prematurely eliminate query-relevant
tokens, depriving the subsequent text-guided stage of critical visual evidence. Our empirical analysis shows that incorporating query guidance into first-stage
pruning better preserves task-relevant evidence and consistently improves performance over
vision-only saliency-based
pruning. We further find that high-variance
attention heads are more sensitive to the textual query and yield more discriminative text-to-
vision attention signals for second-stage
pruning. Motivated by these findings, we propose TReVS, a training-free framework that combines textual relevance with
vision-encoder saliency for pre-LLM
pruning and leverages high-variance
attention heads to remove task-irrelevant
tokens at shallow-to-intermediate layers of the LLM. On LLaVA-1.5-7B, TReVS retains 92.8% of the unpruned baseline performance while
pruning 94.4% of visual
tokens, outperforming prior state-of-the-art methods.
Read the full abstract ▾
All authors Jing Wang, Zhiping Wu, Dongdong Ren, Youfang Han, Wei Zhao, Wenbin Li
Artificial Intelligence · Cross-listed from cs.CR
Concealing LLM-Based Multi-Agent Topology via Phantom Structure Injection
Longzhu He, Zelang Wen, Xinfeng Li et al. (+2 more)
Driven by the rapid advancement of large
language models (LLMs), LLM-based multi-
agent systems (MAS) have emerged as a powerful paradigm for collaborative
reasoning over complex tasks. A key design element of MAS is the communication
topology, which governs information flow among
agents and often encodes proprietary knowledge about the system architecture. However, recent work has shown that such
topologies can be inferred even in black-box settings by exploiting semantic dependencies in observable
reasoning traces, posing significant risks of intellectual property leakage and exposure of system vulnerabilities. To address this threat, we propose MIRAGE, a
topology-concealment framework that preserves the genuine communication
topology for task execution while shaping adversary-facing semantic evidence toward a carefully constructed phantom
topology. Specifically, MIRAGE operates in three stages: (1) phantom
topology synthesis, (2) semantic edge realization, and (3) protected MAS execution. It constructs a phantom
topology structurally distinct from the genuine one, materializes phantom edges as plausible semantic dependencies, and suppresses source-specific cues that could reveal genuine edges absent from the phantom
topology. Extensive experiments across three
topology optimization frameworks and four
benchmark datasets demonstrate that MIRAGE substantially reduces the effectiveness of
topology inference attacks while largely preserving the task utility of the protected MAS.
Read the full abstract ▾
All authors Longzhu He, Zelang Wen, Xinfeng Li, Sen Su, XiaoFeng Wang
Artificial Intelligence · Cross-listed from cs.RO
Risk-Aware Semantic Grounding for Trustworthy LLM-Based Robot Planning
Łukasz Sobczak, Nur Keleşoğlu and Sławomir Piotr Nowak
Large
language models (LLMs) are increasingly used as high-level planners in
robot navigation, but their outputs may become unreliable when instructions are ambiguous, unsupported by the environment, or semantically inconsistent. This paper presents a Risk-Aware Semantic Grounding framework for trustworthy LLM-based
robot planning. Unlike existing LLM-based planners that primarily optimize plan generation, we formulate semantic grounding reliability as a multi-dimensional risk estimation problem. The proposed architecture explicitly models grounding uncertainty through ambiguity,
hallucination and semantic-conflict risks before planning occurs, enabling the system to decide whether to execute the instruction, request clarification, or reject it. To evaluate the approach, we introduce TRUST-NAV, a
benchmark containing both standard
navigation tasks and risk-inducing instruction scenarios. Experimental results show that while conventional LLM planners achieve strong performance on valid
navigation tasks, the proposed framework substantially improves ambiguity detection and semantic conflict rejection. These findings suggest that trustworthy
robot planning should be evaluated not only by task completion, but also by the ability to recognize when execution should not occur.
Read the full abstract ▾
Artificial Intelligence
How Can Recommendation Feedback Evolve Agent Memory?
Shanwen Mao, Mingming Li, Hao Zhang et al. (+4 more)
Content-generation
agents continuously receive impressions, clicks, conversions, and negative feedback from recommendation systems, providing real-world outcome signals for
memory evolution. However, these signals are delayed and noisy, confounded by audience composition, placement, and recommendation policies, and may result from the combined influence of multiple
memories, making accurate attribution difficult. Existing methods rely primarily on immediate feedback or semantic
retrieval and therefore struggle to reliably translate recommendation outcomes into
memory fitness. To address this challenge, we propose TIDE (Trajectory-Informed Directed
Memory Evolution), an external
memory evolution framework driven by delayed recommendation feedback. We further introduce
Memory Evolution Gain (MEG), which measures the utility improvement of evolved
memory over a no
memory baseline on strictly future tasks. TIDE treats
memory as a capacity-constrained population of experiences: temporal and semantic credit assignment estimates contextual fitness, while responsibility credit distributes outcome signals according to the
memories referenced during generation. These signals are then used to reinforce, crossover, mutate, or evict
memories. On an e-commerce membership marketing content-generation
agent, TIDE achieves a +7.75-percentage-point MEG in offline temporal replay and significantly improves both unique click-through rate (UCTR) and activation rate in an online A/B test. On a delayed-label
benchmark, TIDE achieves the lowest mean absolute error (MAE) and root mean squared error (RMSE) and the highest MEG among the compared methods, demonstrating its effectiveness.
Read the full abstract ▾
All authors Shanwen Mao, Mingming Li, Hao Zhang, Zhiheng Li, Yige Wang, Penghua Yu, Junxiong Zhu
Artificial Intelligence
SkillGym: Training Skill-Use Agents with Automatic Verifiable Environment Generation
Renxi Wang, Mingshan Hee, Fajri Koto et al. (+2 more)
Skills equip LLM
agents with professional knowledge and guidance to complete long-horizon and complex tasks. Although skills have been widely adopted in recent
agent paradigms and harnesses, how to synthesize reliable training data and how to train
agents for skill use remain underexplored. In this work, we propose SkillGym, an automatic pipeline to build verifiable environments, collect trajectories, and train skill-use
agents. SkillGym first crawls a large volume of skills from the internet, then keeps those whose workflows can run reproducibly offline. A builder-reviewer pipeline is used to construct difficulty-controlled tasks, spanning four task types, each with a reference solution and an executable verifier. With this pipeline, we build 6.8k environments and collect 19k verified successful trajectories for supervised finetuning. Finetuning on these trajectories improves LLMs of different families and sizes, from 2B to 122B parameters across four skill-use
benchmarks; Our Qwen3.5-9B SFT model outperforms the 397B untrained model on two of them. Further analysis shows that training teaches
agents to invoke skills, raising the rate of reading the relevant skill from 28% to 96%, and that the gains hold across
reasoning structures, extending to task types that form a minority of the training data and to skills held out from training
Read the full abstract ▾
All authors Renxi Wang, Mingshan Hee, Fajri Koto, Timothy Baldwin, Haonan Li
Artificial Intelligence · Cross-listed from cs.DC
DScale: Scaling Block-Diffusion Speculative Decoding with Adaptive Verification
Rongjian Chen, Minxian Xu, Zhengxin Fang et al. (+2 more)
Growing large
language model applications demand efficient inference. At high concurrency, block-
diffusion speculative decoding suffers from verification padding, rejected candidates, and incompatibility between variable prefixes and fixed-shape graphs. Uniform truncation sacrifices acceptable
tokens. We present DScale, preserving drafter architecture, weights, and full draft length. A separate 112K-parameter predictor requires neither confidence calibration nor hardware speed-curve preparation. Path-aware tiles reduce padding. Dynamic verify-length (DVL) allocation packs scored prefixes into half the native verification capacity. Fixed-address workspaces propagate changing boundaries through verification and acceptance while reusing captured graphs. On A100-40GB with tensor parallelism 1, Qwen3-8B and Qwen3-4B cover four
datasets and concurrency 8-32, reusing each target's frozen predictor.
Geometric-mean throughput gains across these configurations are respectively 43.9% and 48.8% over DFlash, 22.2% and 37.7% over DSpark, and 24.4% and 32.0% over Domino, with lower request latency. Cumulative ablations show that adding the three mechanisms successively increases
geometric-mean throughput, while budget adjustment improves accepted-
token retention. GPU profiling shows that complete decode-step time on GSM8K decreases by 30.8-52.5% relative to DFlash
Read the full abstract ▾
All authors Rongjian Chen, Minxian Xu, Zhengxin Fang, Kejiang Ye, Chengzhong Xu
Artificial Intelligence · Cross-listed from cs.RO
Video2STL: Grounding VLM-Generated Temporal Specifications for Robot Learning
Merve Atasever, Keyan Azbijari, Cagan Bakirci et al. (+6 more)
Video-based policy learning is particularly promising, as it illustrates target behaviors without requiring action annotations or embodiment-matched demonstrations. A central challenge is deciding what information should be transferred from the
video to the
robot. Existing approaches commonly convert visual observations into scalar similarity or value signals, or ask foundation models to directly generate reward code. These approaches can make the temporal structure of a task difficult to inspect, ground, and reuse. We present
Video2STL, a framework that converts observation-only
videos into parametric Signal Temporal Logic (STL) specifications and uses the resulting formal representation for
robot learning. A
vision-
language model extracts an embodiment-independent semantic event trace and constructs a bank of symbolic temporal specifications. The model determines the task structure, while numerical predicate thresholds and temporal bounds are grounded from successful
robot trajectories. For policy learning, we separate short- and long-timescale temporal information: short-horizon specifications provide dense rewards through rolling-window quantitative robustness, while a causal monitor over a retained long-horizon specification provides one-time progress rewards for valid temporal prefixes. The same representation supports cross-embodiment transfer from human or animal
videos to
robot control. Across four manipulation tasks,
Video2STL achieves $85.8\%$ average success-once and $67.0\%$ success-at-end, compared with $81.5\%/59.5\%$ for native dense PPO and $65.0\%/42.3\%$ for Text2Reward; in quadruped locomotion, Qwen-3.8 and GPT-5.6-based
Video2STL policies achieve $100\%$ success across velocities from $0.3$ to $2.1\,\mathrm{m/s}$ while remaining competitive in high-speed energy efficiency. Project webpage: \href{https://
video2stl.github.io/}{
video2stl}.
Read the full abstract ▾
All authors Merve Atasever, Keyan Azbijari, Cagan Bakirci, Bo-Ruei Huang, Tolga Izdas, Zahra Shahrooei, Richard Yang, Erdem Biyik, Jyotirmoy V. Deshmukh
Artificial Intelligence · Cross-listed from cs.LG
Hyunjong Ok, Seunggu Kang and Jaeho Lee
Thorough evaluation of
vision-
language models (VLMs) has become prohibitively expensive, as
benchmarks span an ever-broader spectrum of capabilities and new models arrive at a relentless pace.
Benchmark compression methods that preserve model rankings at a fraction of the cost are well studied for
language models, but for VLMs the question remains under-explored. We present PRIMEBench (
Pruning Redundant Items for
Multimodal Evaluation), a
vision-aware hierarchical
benchmark compression framework that substantially reduces evaluation cost while preserving model rankings. This hierarchical framework operates in four stages: data cleaning to remove items answerable without the image and all-correct items, category representative selection to pick one
benchmark per capability category, item
pruning with
Vision-Aware Variance (VAW), and category-count
pruning. VAW combines inter-model variance with a
vision-dependence score computed from
multimodal embeddings alone, while encouraging coverage of diverse items within each
benchmark. On models held out from item selection, it has the highest mean fidelity at the released 5% retention. The hierarchical design lets practitioners stop at any stage to match their compute budget; the released suite removes over 97% of items while preserving model rankings. Beyond compression, our analyses show how VLM evaluation behaves as model panels grow and evolve, providing guidance for designing future
benchmarks that are more efficient, robust to model turnover, and explicit about the limits of evaluation-side
pruning.
Read the full abstract ▾
Artificial Intelligence
Demistifying Data and Simulator Assumptions in Supervised Causal Discovery
Pingchuan Ma, Rui Ding, Bojun Huang et al. (+1 more)
Supervised causal discovery learns to infer causal structure for a new
dataset from training
datasets paired with structural labels. These training pairs are typically simulated, making the simulator both a source of supervision and a carrier of assumptions about causal graphs, mechanisms, and noise. Understanding the resulting predictions therefore requires examining how these assumptions supplement the information available in observational data, which may be compatible with multiple causal graphs. This paper examines that relationship across representative methods available through June 2026. We organize these methods by prediction target, prediction granularity, encoder, structural decoder, and training regime to relate what each method predicts to how it uses data and simulator-based supervision. Using this framework, we distinguish two questions: whether the target is identifiable under the assumed model class, and whether a trained predictor generalizes beyond its training distribution. Restrictions on mechanisms and noise can make otherwise ambiguous causal directions identifiable, but predictive accuracy under those restrictions does not establish transfer when they change. This distinction motivates evaluation that matches metrics to the identifiable graph target and tests changes in graphs, mechanisms, and noise between training and deployment. Extending such evaluation to real data also requires documenting the external causal evidence and uncertainty behind
benchmark reference graphs. Together, these analyses guide method comparison and identify open questions in transfer, test-time adaptation, and uncertainty assessment.
Read the full abstract ▾
All authors Pingchuan Ma, Rui Ding, Bojun Huang, Shuai Wang
Artificial Intelligence · Cross-listed from cs.CL
Learning to Retrieve Missing Evidence for Long-Term Memory QA
Yi-Xuan Deng, Yi Zhang, Wei Liu et al. (+2 more)
Long-term
memory enables
language models to use past interactions in future conversations. However, evidence needed to answer a question may be scattered across distant turns, while the question itself omits clues needed to locate it. Retrieved facts can reveal these clues, motivating
retrieval decisions conditioned on evidence already found. We introduce MERA (Missing-Evidence
Retrieval Augmentation), which separates globally searchable
memory from a question-specific evidence state. Verified evidence guides subsequent
retrieval without restricting access to the global
memory. We train a lightweight planner through reinforcement learning, rewarding queries that recover previously missing evidence. MERA achieves strong answer accuracy across Qwen3-30B and GPT-4o-mini backbones. With Qwen3-30B for evidence processing and answer generation, the trained 0.6B planner achieves 77.40% accuracy on LoCoMo and 71.29% on LongMemEval-S, exceeding a 30B planner without
retrieval-grounded training by 4.10% and 3.96%, respectively. On LoCoMo, later
retrieval rounds increase cumulative evidence recall from 55.5% to 80.5%.
Read the full abstract ▾
All authors Yi-Xuan Deng, Yi Zhang, Wei Liu, Chao Xue, Shuojin Yang
Artificial Intelligence · Cross-listed from cs.CV
LazySloth: Bounded LLM-based Lazy Tree Search for Fast Long Video Comprehension
Arka Mukherjee, Kaleen Shrestha, Larissa Zhu et al. (+1 more)
Modern
vision-
language models (VLMs) have shown promising results in long-
video understanding due to the rich semantic information they can capture. However, most methods focus on coarse captioning of extracted image frames that are computationally inefficient and require models with large context windows. While past work has explored efficient methods through
multimodal retrieval-augmented generation (RAG), they rely on lossy
embeddings that lose temporal context and fine-grained detail. Few works to date have investigated how VLM-based query-relevant information
retrieval can be optimized. We introduce LazySloth, an efficient tree-based search method that speeds up
video comprehension and
retrieval tasks 2.9-8.3x (compared to existing agentic methods) through bounded captioning of portions of the
video considered irrelevant by a VLM of the
video. Compared to contemporary specialized
video-understanding VLMs and RAG-based methods, LazySloth achieved similar or better final task accuracy across two recent open-source base VLMs–Gemma 4 31B and Qwen3.6 27B–across four
benchmarks. LazySloth reduced the gap between the base open-source model and a closed-source model, GPT-4o. Ablations showed that replacing VLM scene understanding with CLIP-based
retrieval cost 8.8-19.9% in accuracy, while lazy tree construction matches eager construction at a fraction of the captioning cost. With LazySloth, we demonstrate the possibility of faster long-
video comprehension without substantial loss in performance.
Read the full abstract ▾
All authors Arka Mukherjee, Kaleen Shrestha, Larissa Zhu, Maja Matarić
Artificial Intelligence · Cross-listed from cs.LG
Simultaneous Neural Optimal Transport
Milena Gazdieva, Kirill Sokolov, Jiawei Chen et al. (+2 more)
Optimal Transport (OT) provides a principled framework for learning transformations between probability distributions from unpaired samples. In many applications, however, a single transformation must map several source distributions to a common target distribution. For example, image restoration might require handling different types of degradation without knowing the degradation of each input at inference time. Simple approaches of pooling the source distributions only encourage
alignment with the target at the aggregate level and may leave individual sources misaligned. In our paper, we consider the simultaneous OT problem which formalizes the task of learning a shared transport map that minimizes the average transport cost while aligning each source distribution with a prescribed target. We propose a neural method for solving the simultaneous OT problem by learning a shared transport map that minimizes the average transport cost while aligning each source distribution with a prescribed target. We derive a max-min formulation for learning this map. We illustrate its application to image restoration, where a single model handles multiple degradation types using a common collection of clean target images.
Read the full abstract ▾
All authors Milena Gazdieva, Kirill Sokolov, Jiawei Chen, Evgeny Burnaev, Alexander Korotin
Artificial Intelligence · Cross-listed from cs.SE
Complexity-Aware Evaluation of LLM Comprehension
Ali Mohammadi Esfahani, Nafiseh Kahani and Samuel A.Ajila
Large
language models (LLMs) are increasingly used for software engineering tasks that require understanding existing source code, including behavior prediction, function explanation, debugging, and code review. However, aggregate
benchmark accuracy can conceal how model reliability changes as source code becomes structurally more complex. This paper presents a complexity-aware framework for evaluating LLM code comprehension using cyclomatic complexity, nesting depth, branching factor, and Halstead volume. We evaluate DeepSeek-Coder-V2 and Llama through two complementary tasks: automatic input-output prediction over 300 Python functions and manually assessed semantic comprehension over a balanced subset of 60 functions. The functions are grouped into Low-, Medium-, and High-complexity bands. DeepSeek-Coder-V2 achieves an overall automatic accuracy of 78.33%, compared with 70.33% for Llama. However, accuracy decreases substantially from Low to High complexity, from 93.52% to 52.78% for DeepSeek-Coder-V2 and from 87.04% to 47.22% for Llama. Incorrect predictions are consistently associated with higher values of all four complexity metrics, and correlation and logistic-regression analyses confirm broadly comparable negative associations between structural complexity and correctness. Manual semantic comprehension shows the same degradation pattern, with accuracy decreasing from 100.00% to 75.00% for DeepSeek-Coder-V2 and from 90.00% to 60.00% for Llama. These findings demonstrate that complexity-aware evaluation provides a more diagnostic assessment of LLM code-comprehension reliability than aggregate accuracy alone.
Read the full abstract ▾
Artificial Intelligence
Routing Should Pay for Itself: Sparse Supervision for Economical LLM Routing
Guannan Lai, Gelin Bian, Hao-Xuan Ma et al. (+5 more)
Large
language model (LLM) routing reduces serving cost by assigning each query to an appropriate model while preserving response quality. Learning such a router, however, often requires executing multiple candidate models on historical queries to collect query–model quality feedback, creating a nontrivial supervision cost before deployment. Existing work largely focuses on serving-time efficiency, overlooking whether the resulting savings are sufficient to recover this upfront expenditure. We further observe that routing quality often saturates well before all query–model feedback is collected, suggesting that dense supervision can be economically over-provisioned. We propose SaveRouter, a sparse-supervision routing framework that selectively acquires informative model feedback and shares capability information across related queries, while retaining query-level refinement for fine-grained routing. We evaluate routing by jointly accounting for supervision expenditure and subsequent serving-time savings. Across four routing
benchmarks, the main setting uses only about 33–41% of available training feedback while maintaining competitive or better routing quality, and reduces the break-even deployment volume by approximately 1.9–9.5 times compared with the fastest conventional router. Further analysis shows that acquiring more supervision is not always economically preferable: the supervision level that minimizes serving cost can differ from the one that achieves the earliest payback. Our code is publicly available at https://github.com/LAMDA-Model-Reuse/SaveRouter.
Read the full abstract ▾
All authors Guannan Lai, Gelin Bian, Hao-Xuan Ma, Jun-Peng Jiang, Long Chen, Jian-Dong Liu, Zhi-Hao Tan, Han-Jia Ye
Artificial Intelligence
Direct Experience World-Model Optimization: Learning the World Beyond Action Imitation
Xiangcheng Zhan, Zirui Chen, Yicheng Zhao et al. (+2 more)
World-Action Models (WAMs) couple action generation with predictions of how physical interactions unfold. However, current post-deployment learning paradigms typically improve behavior without requiring better world predictions. Especially in dexterous manipulation, small execution errors can compound in high-dimensional action spaces, hindering policy improvement and pushing interactions beyond the world model's training distribution. Motivated by this, we propose Direct Experience World-Model Optimization (DEWO), a post-deployment learning paradigm for WAMs that, alongside action imitation, refines world representations through visual experience to better condition action generation. Specifically, it identifies interaction turning points and learns from successful and failed futures to support classifier-free guidance. An additional value head estimates task progress from
video representations and activates guidance when progress stalls during inference. Across five DexJoCo tasks, DEWO improves average success across all three WAM formulations. Ablations show that visual supervision from successful and failed continuations improves both prediction and control beyond action supervision alone. On four real-world tasks across Wuji and Sharpa, 3 x 3 grid evaluations show that two rounds of deployment learning increase success from 51.0% to 71.7% in
cells with at least one initial success, a gain of 20.7 percentage points. These findings support continued predictive learning for improving control through deployment experience, making world modeling an active part of WAM adaptation.
Read the full abstract ▾
All authors Xiangcheng Zhan, Zirui Chen, Yicheng Zhao, Ziteng Gao, Shuo Yang
Artificial Intelligence · Cross-listed from cs.CV
Do-JEPA: From Masking to Intervention in Latent World Models
Hossein Resani and Javen Qinfeng Shi
Latent world models are trained to predict what happens next, so nothing in their objective separates what an action caused from what merely co-occurred with it. Object-masking models such as C-JEPA intervene on what the predictor can see; we intervene on what physically happens. From one saved simulator state we run the dynamics under an action $a$ and under a reference action $a_{\varnothing}$, and train the model to predict the difference $\Delta z=z^{a}-z^{a_{\varnothing}}$ between the two latent futures. The resulting objective, Do-JEPA, has an effect loss, a support loss (where the action enters), a propagation loss (where its effect travels) and invariance losses (what must not change). In a synthetic system with object-aligned variables, support supervision finds the directly intervened object in 99.95% of test cases, where a sparse action mask sends the action to a nuisance slot in every case, and response-onset supervision recovers the ring-shaped propagation graph (edge AUROC 0.975 vs. 0.624). From pixels, the effect loss beats a control trained on exactly the same data: it lowers latent effect error by 28.4% on an end-to-end LeWM model and physical effect error by 13.5% when trained and tested on natural action sequences, and on three independently generated CausalWorld
benchmarks it lowers responsive effect error by about 20% under physics shifts and the latent context sensitivity of predicted effects by 66%. Trained from scratch it costs factual accuracy; fine-tuning an existing model with it removes this cost. Together, these results show that intervening on the world, rather than on what the model sees, helps latent world models predict what their actions cause.
Read the full abstract ▾
Artificial Intelligence
Beyond Prompt Count: How Data Shapes Transfer in On-Policy Distillation
Jiaxuan Wang, Jiafei Lyu, Yuchen Cai et al. (+8 more)
On-policy
distillation (OPD) trains students using teacher feedback on their own sampled responses, yet how
prompt choice shapes transfer across teacher-student pairs remains poorly understood. We systematically study
prompt quantity, source, and selection across RL- and SFT-continuation pairs and cross-model settings. We find that OPD can be highly
prompt-efficient: a few
prompts can approach large-pool performance, with four DAPO
prompts matching the observed mathematics score of 3,840 DeepMath
prompts. However,
prompt utility is relational rather than intrinsic: changing only the teacher can reverse the relative effectiveness of mathematics and code
prompts. To characterize these transfer differences, we analyze parameter and functional changes across
prompt supports and model pairs. Functional
alignment with the teacher varies across supports and target tasks; in continuation pairs, teacher-aligned prediction changes can coexist with weak parameter
alignment. Continued OPD on effective supports can restore performance after unfavorable transfer. Finally, targeted selection does not consistently outperform uniform random sampling, and filtering out a source that performs poorly alone yields no consistent gain across three paired support draws. Overall, our results distinguish
prompt efficiency from
prompt interchangeability and show that effective data choice depends on the teacher-student pair and target capability, with random sampling providing a competitive baseline in the studied settings.
Read the full abstract ▾
All authors Jiaxuan Wang, Jiafei Lyu, Yuchen Cai, Siye Wu, Pengyuan Wang, Jiashun Liu, Xiang Cheng, Kai Yang, Yangkun Chen, Saiyong Yang, Lan-Zhe Guo