Peer Revenue Impact factor: volatile

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Computation and Language

cs.CL Watch Computation and Language

Friday 2 October

Computation and Language · Cross-listed from cs.LG

OPD Before RL: Warm-Starting Rubric-Based RL with On-Policy Distillation

Xinpeng Wang, Wei Shi, Yu-Chia Chen et al. (+3 more)

Many useful language-model tasks cannot be evaluated by exact outcome verification. Rubric-based reinforcement learning (RL) addresses this issue by scoring open-ended responses against explicit criteria. However, because the reward is assigned after the complete response, the training signal does not directly identify which individual decisions contributed to the final score. We propose a two-stage training framework that uses rubrics first as privileged teacher context for dense token-level supervision, then as rewards for further RL. In the first stage, rubric-privileged on-policy distillation (RP-OPD), a student without access to the rubric matches a rubric-aware teacher's next-token distributions at student-generated prefixes. In the second stage, RL directly optimizes the rubric reward and improves beyond the observed distillation plateau. We evaluate the framework on health and science tasks using open-weight models. Across HealthBench, ResearchQA, and RubricHub Science, we compare post-training methods and vary the amount of SFT or RP-OPD training before RL, finding that our two-stage framework achieves the highest scores among the methods evaluated. RP-OPD + RL shows limited signs of reward hacking on RubricHub Science, whereas the SFT + RL baseline increasingly receives high rewards for claims of rubric compliance without providing the required content. These findings support using rubrics to guide on-policy distillation before applying rubric-based RL.
Read the full abstract

All authors Xinpeng Wang, Wei Shi, Yu-Chia Chen, Maria Zontak, Yun He, Richard Yuanzhe Pang

Computation and Language

Automatic Evaluation of Mental Health Stigma in Online Communication

Naomi Baes, Jemima Kang, Nick Haslam et al. (+4 more)

Mental health stigma has profoundly harmful impacts but its complexity makes it difficult to evaluate. Stigma may involve explicit derogation, but also subtler forms of blame, fear, paternalistic pity, social distancing, structural exclusion, and discrimination. We introduce a theory-grounded benchmark for automatic evaluation of mental health stigma in online communication, consisting of naturally occurring online news and social media text annotated with a fine-grained taxonomy of stigma across multiple mental health conditions. Our annotation framework comprises a binary stigma-detection task and a multi-level taxonomy covering (i) stigma mode, (ii) domain, and (iii) specific components of certain forms of stigma. We apply this framework to texts mentioning six mental health conditions and evaluate large language models alongside stigma-related classifiers for detecting sentiment, toxicity, and hate speech. Results show that mental health stigma is not well captured by models trained to detect these neighboring constructs, and that LLMs often overpredict stigma unless given explicit operational rules - mirroring the importance of decision rules in human annotation. We release the publicly available part of benchmark, annotations, prototypical exemplars of stigma and code at: https://github.com/jemimakang/mh_stigma.
Read the full abstract

All authors Naomi Baes, Jemima Kang, Nick Haslam, Chris Groot, Alsa Wu, Luc Raszewski, Yulia Otmakhova

Computation and Language

Improving Atomic-Fact Recall via Focused Views in Unstructured Knowledge Editing

Ding Wu, Ye Zhang, Haoyu Wang et al. (+1 more)

Large language models (LLMs) increasingly serve as general-purpose interfaces to factual knowledge, but their parameters do not automatically reflect information that changes after pretraining. Knowledge editing (KE) provides a targeted alternative to costly retraining by modifying selected knowledge and preserving unrelated knowledge and general capabilities. Conventional KE uses structured factual triples, whereas unstructured KE (UKE) uses free-form passages containing multiple facts. Nonetheless, existing UKE editors exhibit a failure mode known as context reliance: edited LLMs can often reproduce the editing passage but fail to reliably recall its individual facts without the original passage context. We identify context-induced difficulty underestimation under the standard passage-level editing objective: later facts receive increasingly rich ground-truth context and consequently incur lower initial losses, making them appear easier to learn. In response, we propose FOVEATED, a plug-and-play framework that constructs focused views of each sentence by randomly shifting the Rotary Position Embedding (RoPE) positions assigned to the keys of its preceding context. The perturbation is applied during editing and removed afterward, leaving the model's native positional encoding unchanged at inference time. We instantiate FOVEATED for both direct-optimization and locate-then-edit editors. We theoretically analyze how FOVEATED counteracts context-induced difficulty underestimation and empirically demonstrate consistent improvements across five KE editors, two LLM backbones, and three benchmarks.
Read the full abstract

All authors Ding Wu, Ye Zhang, Haoyu Wang, Tianci Liu

Computation and Language

AptMQL-Bench: From Text-to-SQL to Text-to-MQL via Access-Pattern Schema Design and Data-Preserving Migration

Hy Nguyen, Nabi Rezvani and Robin Vujanic

Document databases such as MongoDB are core infrastructure for modern applications, and natural-language interfaces to them—text-to-MQL—would let non-experts query complex, semi-structured data without mastering the query language. Progress on this task depends on high-quality benchmarks, which are most practically obtained by converting an existing text-to-SQL benchmark to the document setting. Unfortunately, existing efforts rely on heuristics for mechanical conversion: the document schema mirrors the relational foreign-key graph, and each query mirrors its source SQL. As a result in our experiments, these approaches fail to migrate 6 of 21 BIRD databases outright, silently drop up to 25.9% of rows on others, and yield schemas whose ground-truth queries run over an order of magnitude slower as the data scales. We instead propose a conversion pipeline, driven by coding agents with human-in-the-loop verification, that designs each document schema from expected access patterns and rewrites queries to be MongoDB-native. Applying it to BIRD, we build an access-pattern-based text-to-MQL benchmark (AptMQL-Bench). It includes 21 document-oriented databases, 3,186 natural-language requests, and their associated MQL queries—whose databases are migrated from SQLite without data loss and scale efficiently. The strongest model, Claude Opus 4.5, achieves only 57.38% accuracy without external knowledge evidence and 70.34% with it. This indicates that realistic text-to-MQL generation remains challenging.
Read the full abstract

Computation and Language

When History Fails to Become Experience: Action Calibration in Language Agents

Jingyu Liu, Zhiwen Wang, Yuxin Jing et al. (+2 more)

Language agents should draw on prior attempts and environmental feedback to improve subsequent decisions within the same task. However, providing additional interaction history can sometimes reduce task success, suggesting that agents do not consistently use this information effectively. To investigate this limitation, we examine how agents use history. We find that history improves task completion overall, yet much of this benefit persists even when past actions are shuffled. Disrupting the correspondence between actions and observations causes only a modest decline in task success. We therefore hypothesize that agents do not reliably connect past actions with their outcomes when deciding how to proceed. To test this hypothesis, we explicitly label each returned observation as the outcome of the preceding action. This simple annotation improves task success and reduces next-action repetition without introducing new environmental information. Building on this insight, we introduce a learned calibrator that explicitly reassesses past actions and selectively records experience to guide subsequent decisions, improving task success beyond outcome labeling alone.
Read the full abstract

All authors Jingyu Liu, Zhiwen Wang, Yuxin Jing, Huanyu Zhou, Yong Liu

Computation and Language

EpiWorld: Grounding LLM Policy Agents in Epidemiological World Models

Zeeshan Memon, Yiqi Su, Kai Shu et al. (+2 more)

Epidemic intervention policies are textual artefacts that human decision-makers interpret, justify, and revise through natural language, making large language models a natural candidate for epidemic policy reasoning. A naive LLM, however, lacks the epidemic dynamics needed to project intervention consequences, the quantitative surveillance signals required to assess severity, and the institutional constraints that define admissible actions. We present EpiWorld, a closed-loop framework that grounds an LLM policy actor in a learned action-conditioned epidemiological world model and a tiered skill library of public-health protocols, surveillance tools, and adaptive lessons accumulated through after-action analysis. Given a candidate intervention, the world model predicts regional epidemic evolution and enables fast counterfactual rollouts that provide feedback for policy selection and refinement. Outcomes of simulated futures are distilled into reusable lessons while protocol constraints remain fixed, allowing the decision process to improve without sacrificing interpretability or controllability. We evaluate both the world model and the end-to-end framework on retrospective COVID-19 and Influenza datasets: the world model achieves the best out-of-distribution Peak-MAE among all forecasting baselines, and the closed-loop framework reduces cumulative hospitalisation by up to 59% across datasets and by an average of ~16% across six LLM backbones, outperforming reinforcement-learning and optimal-control policy baselines.
Read the full abstract

All authors Zeeshan Memon, Yiqi Su, Kai Shu, Naren Ramakrishnan, Liang Zhao

Computation and Language

Beyond Correctness: Resolving Underspecification in Agentic Text-to-SQL

Wen-Zhi Li, Yue Gong, Konstantinos Kanellis et al. (+1 more)

Agentic Text-to-SQL systems can interact with users to clarify underspecified queries before generating SQL. However, a correct execution result does not necessarily imply that the agent has adequately resolved the underlying underspecification: the agent may silently make unverified assumptions that happen to match the intended answer. We show that this behavior is driven in part by premature clarification termination. Although forcing an agent to ask more questions improves execution accuracy, ambiguities are concentrated in earlier interactions, making brute-force questioning inefficient. More importantly, even when explicitly prompted to plan its clarification process, the agent frequently abandons questions that it has already identified as relevant. To address this failure mode, we introduce PlanPool, which externalizes the clarification plan as a mutable question pool. Every planned question must be explicitly asked or dropped before submission, while newly discovered ambiguities can be added during interaction. Across three benchmarks derived from BIRD-Interact and Spider, PlanPool consistently improves ambiguity coverage and reduces silent failures over unconstrained and prompt-based alternatives, while maintaining competitive execution accuracy. Our results highlight an important distinction in agentic reasoning: identifying missing information is not sufficient, and the agent must also reliably maintain and resolve it before committing to an answer.
Read the full abstract

All authors Wen-Zhi Li, Yue Gong, Konstantinos Kanellis, Balakrishnan Murali Narayanaswamy

Computation and Language

TPBench: A Turning-Point Benchmark for Dialogue Compression

Minji Park, Seunghyun Yoon and Hyuk Lim

A compressor can keep the facts of a dialogue and still drop the turn that changed them. A user corrects a price, reverses a choice, or adds a constraint. We call this failure turning-point eviction. One overall retention score hides it, because that score mixes what the user first wanted with what the user wants now. We introduce TPBench, which evaluates three complementary information targets at shared nominal retention budgets. P1 asks for the user's initial goal. P2 asks for the current value of a slot the user revised. P3 asks for both, in dialogues with a late annotated slot update. The current-value answers come from the human dialogue-state annotations of MultiWOZ and SGD. The initial-goal answer is the first sentence of the first user turn. Neither requires new crowdsourcing. The probe-specific evaluations rank compression methods differently. On the joint probe at a retained fraction of 0.30, every tested compressed method remains below full context with the main Llama reader. Deleting the turn that carries the update sharply lowers current-value accuracy, while deleting one matched irrelevant turn leaves it unchanged. A Mistral reader repeats the P2/P3 rankings and the joint-probe gap. Current-value recovery is tested on an additional corpus, LongMemEval-KU, and on Chinese RiSAWOZ: full context has the highest accuracy, and recency has the highest compressed-method mean in both evaluations.
Read the full abstract

Computation and Language

WakeKV: Reactive, Reversible KV Residency for Heads That Change Their Minds

Utkarsh Ranjan

Most KV-cache compression methods classify attention heads once, either offline or during prefill, and keep this classification fixed throughout generation. Across three models (1.5B-8B) and three regimes (needle retrieval, long chain-of-thought, and multi-turn recall), we measure head behavior on four model-regime combinations and find that most heads change their reading behavior at least once during generation. We introduce WakeKV, a reactive residency policy that moves cooling heads to a recoverable CPU reservoir rather than freezing or permanently evicting their state. At matched memory or budget, WakeKV consistently improves miss rate over frozen classification and destructive eviction, evaluated across five model-regime combinations and over three cited baselines (SnapKV, uniform R-KV, and ReasonAlloc) across four eligible combinations. A FlexiCache/vLLM implementation on Mistral-7B confirms the benefit on real hardware, improving throughput while retaining LongBench quality.
Read the full abstract

Computation and Language

Silent Dissent: LLM Agents That Yield to the Majority Still Represent Their Original Premise

Ziang Ni and Peng Zou

Multi-agent debate is increasingly used to reach consensus among LLM agents, yet agents often yield to a unanimous majority. When an agent changes its answer, has it changed its mind or only its statement? We study this with two-hop factual questions whose intermediate entity (the bridge, e.g. the country in "the capital of the country where the Sagrada Familia is located") is never stated by anyone. Scripted peers, in the role of Asch's confederates, unanimously assert a wrong answer taken from another fact with a different bridge. At the moment the agent answers, we read the bridge from its residual stream with the Jacobian lens (J-lens) and, for comparison, the logit lens. In pre-registered tests on held-out facts with four open-weight models, agents of Qwen3.5-4B, Qwen3.6-27B and Gemma-4-E4B-it that gave in still represented their original bridge in the pre-registered layers below the output (hit@100 above a control entity: 0.85, 0.22 and 0.24), where the logit lens rarely ranked it among the top 100 tokens (0.00-0.06). These agents also represented the bridge behind the peers' answer, beyond a mention baseline. A pre-registered addendum hid the agent's earlier answer or removed it: agents that gave in still represented their original bridge in all four models (0.43, 0.29, 0.37 and 0.25 with the answer hidden), including Llama-3.1-8B-Instruct, which barely did so with its answer in view (0.03). The premise can thus be computed from the question alone while the agent states the majority's answer. Hiding the earlier answer also changed conformity: Qwen3.5-4B gave in on 89% of questions instead of 8%. In exploratory interventions, injecting the bridge's J-lens direction brought agents back to their original answer only in the two Qwen models. Stated consensus in multi-agent debate can thus overstate agreement. We also report the negative results of our pre-registered program.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

Learning from Evolving Errors: Adaptive Iterative Repair for On-Policy Distillation

Rui Li, Liyang He, Zheng Zhang et al. (+3 more)

On-policy self-distillation (OPSD) supplies dense token-level feedback on trajectories sampled from the student's own policy, a richer training signal than the outcome-level rewards of reinforcement learning. This feedback comes from a teacher conditioned on a full reference solution unavailable to the student. The reference solution specifies the target but not how to move from the student's current error toward it, creating a solution-conditioned shortcut risk. We introduce AIR-OPD, an adaptive iterative repair framework for on-policy distillation that provides error-to-repair supervision. Given a failed response, a guidance generator synthesizes repair guidance for the current error. The student samples an on-policy retry with this guidance. If the retry remains incorrect, the generator produces new repair guidance for the newly observed error. At each round, a fixed teacher receives the guidance as privileged context and supervises the student on an error-aligned region of its latest failed response. Outcome-aware stage weighting favors early repair stages and credits stages whose immediate retry passes verification. We train AIR-OPD on the DAPO-Math-17K dataset and evaluate on AIME24, AIME25, and HMMT25, alongside out-of-distribution tests on MMLU-Pro and GPQA. We examine two guidance sources, self-guidance from the current student policy and external guidance from a larger model. For both Qwen3-4B and Qwen3-8B, AIR-OPD attains the best mathematical-reasoning averages, improving over the strongest baseline by up to 3.6 points, while preserving base-model performance on the out-of-distribution benchmarks.
Read the full abstract

All authors Rui Li, Liyang He, Zheng Zhang, Zhenya Huang, Linbo Zhu, Qi Liu

Computation and Language · Cross-listed from cs.AI

Large language models exhibit unreliable updating of clinical judgment as patient evidence evolves

Min Zeng and Rui Zhang

Large language models (LLMs) are increasingly explored for clinical reasoning, but whether they appropriately revise judgments as patient evidence evolves remains unclear. We evaluated longitudinal belief updating using matched intensive-care trajectories from electronic health records. Across diverse LLMs, conditioning on a preceding judgment more often increased than reduced prediction error when estimates changed, replicated for a second endpoint. Controlled interventions revealed two failure modes. First, with preceding assessment fixed, models responded more strongly to worsening than matched improving respiratory evidence; this asymmetry persisted after headroom normalization at moderate and strong evidence levels. Second, with current evidence fixed, increasing prior risk from 10% to 90% shifted estimates by 26.2 percentage points, demonstrating causal influence of prior model beliefs. Prompting did not restore reliable updating. Evidence-Validated Longitudinal Update (EVLU) identified fewer, more reliable revisions, revealing a reliability-coverage trade-off. These findings establish longitudinal belief updating as a distinct dimension of LLM reliability.
Read the full abstract

Computation and Language · Cross-listed from cs.HC

Asterism: Exploring and Synthesizing Scattered Observations into Literature-Grounded Hypotheses and Theories

Joseph Chee Chang, Michael D'Arcy, Amy X. Zhang et al. (+15 more)

A theory draws many independent observations into one framework with novel hypotheses. A researcher building such a theory must synthesize observations scattered across many papers, each describing related concepts but often in different terms. Which concepts matter most also depends on their preferences and research questions. Recent approaches scale theory synthesis with LLMs, but automate away choices and intuitions from researchers. We present Asterism, which extracts observations from hundreds of papers as concept-relation triples, with concepts unified in a hierarchical ontology. Researchers curate an evidence graph using the ontology and aggregate observations at different levels of granularity to focus theory formation on specific phenomena of interest. In a field deployment (n=10), researchers worked from observations to theories, and kept concepts and hypotheses fitting their preferences. In two case studies, teams of immunology and agriculture researchers discovered mechanisms outside their standard analyses and constructed hypotheses worth follow-up experiments.
Read the full abstract

All authors Joseph Chee Chang, Michael D'Arcy, Amy X. Zhang, Pao Siangliulue, Sangho Suh, Aakanksha Naik, Jena D. Hwang, Javier Ramos Benitez, Stella Wroblewski, Matt Latzke, Michael Cuoco, Ruben Lozano-Aguilera, Kris Ganjam, Joel Chan, Doug Downey, Peter Jansen, Kyle J. Travaglini, Daniel S. Weld

Computation and Language · Cross-listed from cs.LG

LEAP: Learning Efficient Action Proposals For LLM Agents

Zhen Xu, Qizheng Zhang, Gerry Wan et al. (+2 more)

LLM agents are known to be slow in rollouts. An agent completes a task one step at a time. At each step, it reasons and then chooses an action to execute. The next step and action cannot start until the previous one has finished. Speculative decoding accelerates the rollouts at the reason phase by drafting and verifying the inference tokens. Recent works have also started to apply similar ideas at the action phase. These works use off-the-shelf models, usually large, to draft action proposals for target model to verify. Large drafters match the target more often but take longer to propose, while small off-the-shelf models are fast but rarely make the same decision as the target. We ask a more general question: what determines the end-to-end speedup of action speculation? To answer it, we develop a latency framework for the speculative round. The framework compares what a round gains with what it costs. The gain depends on how well the drafter predicts the target and on how many steps the task can take before it ends. The cost comes from drafting, from waiting for target verification and from executing tools. Guided by the framework, we introduce LEAP (Learning Efficient Action Proposals) which keeps the drafter small and makes it accurate by training it on the target actions sequences. With a small 0.6B model, LEAP agrees with the target on most decisions and makes agents up to 60% faster in end-to-end wall clock time, with no systematic change in task success. Across various datasets, target models and draft models, the framework accounts for most of the measured speedups. We also show the draft model can be online trained with no prior trace collection and match the performance of offline training, making LEAP practical to deploy in the real world.
Read the full abstract

All authors Zhen Xu, Qizheng Zhang, Gerry Wan, Shang Zhu, Ce Zhang

Computation and Language

Large Language Continuous Diffusion Models

Zhihan Yang, Wei Guo, Jean-Marie Lemercier et al. (+14 more)

Despite the success of discrete diffusion language models (dLMs) for fast parallel decoding, their non-smooth, high-dimensional space hinders trajectory steering for reasoning and inference acceleration. To overcome this, we present Sigma, the first large-scale (3B/8B) continuous dLM built on steerable, low-dimensional ODE/SDE latent trajectories. Trained blockwise via likelihood optimization, Sigma jointly denoises Gaussian-corrupted token embeddings while learning an optimal embedding geometry. To accelerate training, Sigma leverages pre-trained weights from autoregressive (AR) models for warm-starting. During inference, we identify classifier-free guidance and score temperature as essential for high-fidelity reasoning and coding. Across comprehensive math reasoning and coding evaluations against state-of-the-art discrete counterparts (masked dLMs and AR baselines), Sigma achieves competitive performance with discrete models on standard benchmarks (e.g., GSM8K, Minerva, HumanEval, MBPP) after pre-training and on challenging reasoning tasks (e.g., MATH-500, AIME) after supervised fine-tuning. Beyond performance parity, we uncover key structural properties unique to continuous dLMs: (i) embedding-space steering effectively governs the quality-diversity trade-off, yielding strong pass@k performance and (ii) continuous trajectories enable graceful degradation for low NFEs and efficient distillation. These establish continuous dLMs as a promising paradigm for efficient language generation.
Read the full abstract

All authors Zhihan Yang, Wei Guo, Jean-Marie Lemercier, Simon Welker, Yonggan Fu, Mohammad Mahdi Kamani, Sajad Norouzi, Julius Berner, Tomas Geffner, Karsten Kreis, Yongxin Chen, Molei Tao, John Thickstun, Pavlo Molchanov, Ante Jukić, Arash Vahdat, Morteza Mardani

Computation and Language · Cross-listed from cs.AI

VERSE: Verified Self-Evolving Optimizer for Agent Harnesses

Zekai Wang, Yingqiang Ge, Zekun Wang et al. (+6 more)

Harness evolution improves an LLM agent's prompts, tools, and workflow, while the optimizer's own tools and procedures often remain fixed. We study whether an optimizer can improve another agent more effectively by also improving how it diagnoses failures, develops edits, and tests their effects. Two observations guide our design. In a controlled study, optimizer self-evolution fails to improve performance without execution-based verification, but achieves the best result of that study when verification is available. Across five executors, self-evolving optimizers build their own tools for failure analysis, verification, training audits, and workflow control. Motivated by these findings, we introduce VERSE, a Verified Self-Evolving optimizer for agent harnesses. VERSE lets the optimizer test draft edits, replay failures, and perturb suspected steps before submission, while tracking fixes and regressions across rounds. Using this feedback, the optimizer revises both the executor harness and its own prompts, skills, tools, hooks, and notes, while the weights of the optimizer and executor models stay fixed. Under a shared protocol with disjoint training, validation, and test tasks, VERSE improves all four evaluated harness optimizers on held-out SWE-rebench tasks and newer out-of-distribution tasks in five languages. Its best validation-selected harness reaches 42.3% and 37.7% accuracy, respectively, against 39.2% and 29.3% for the strongest baselines. Code is available at https://github.com/wzekai/VERSE.
Read the full abstract

All authors Zekai Wang, Yingqiang Ge, Zekun Wang, Hai Wang, Yuhui Xu, Joshua Frandsen, Shancong Fu, Ashia C. Wilson, Chandan K. Reddy

Computation and Language

Learning When to Commit from Partial Speech for End-to-End Simultaneous Speech Translation

Hieu Hoang and Amittai Axelrod

Simultaneous speech translation must emit useful target text before the source is complete while preserving every committed token. We adapt a full-utterance speech language model using prefix supervision derived from its own complete- and partial-waveform translations, requiring neither transcripts nor human translations. We compare single-turn forced-prefix and multi-turn append-only decoding, use a confidence threshold to control the inference-time quality–latency trade-off, and vary the density of training prefixes with a separate synthesis margin. On FLEURS and CoVoST2 in three language directions, prefix training improves quality–latency frontiers over the unadapted model, and confidence provides the broadest consistently competitive operating range. Multi-turn decoding is generally stronger at low latency; under multi-turn training, commit-calibration error falls by 63–68% overall and 68–80% at early prefixes, whereas single-turn training provides only modest overall calibration gains and no early-prefix improvement. A small synthesis margin sometimes extends the frontier to lower latency, particularly on shorter utterances, while a larger margin degrades translation quality and calibration. Prefix adaptation therefore improves simultaneous speech translation, especially under multi-turn append-only decoding, while synthesis density introduces a non-monotonic quality–latency trade-off.
Read the full abstract

Thursday 1 October

Computation and Language

Exploring the Design Space of Permutation Diffusion LLMs

Zhe Xu, Mingfu Liang, Kavosh Asadi et al. (+16 more)

Permutation generation with language models requires both text-conditioned ranking and a global one-to-one assignment between candidates and output positions. We study this problem through three diffusion-LLM parameterizations: hard-assignment prediction over shuffled candidate-index sequences, soft-assignment prediction over doubly stochastic matrices, and direct ordering-score prediction followed by sorting. The three families place permutation constraints at different stages of prediction and decoding and therefore support different training objectives and sampling procedures. We evaluate these designs on recommendation reranking across three Amazon domains with candidate sets of size 10 and 20. At one function evaluation, the hard- and soft-assignment variants achieve Recall@1 comparable to a constrained masked-diffusion baseline, whereas direct ordering-score prediction guarantees valid permutations without Hungarian assignment or Sinkhorn normalization. Additional function evaluations do not consistently improve ranking quality, and their effect depends strongly on the state representation and training objective. These results show that corruption, prediction parameterization, objective, and decoder should be designed jointly for permutation diffusion with language-model denoisers.
Read the full abstract

All authors Zhe Xu, Mingfu Liang, Kavosh Asadi, Jean-Pascal Billaud, Tianyi Chen, Yanhui Guo, Christina Huang, Baekjin Kim, Jimmy Kim, Yufei Li, Daqi Sun, Jiachen Sun, Dian Yu, Zaiwei Zhang, Kaushik Rangadurai, Frank Shyu, Sandeep Pandey, Xi Liu, Luke Simon

Computation and Language

Following the Flow the Law Requires: Training Voice Agents to Transfer to Unseen Regulated Procedures in Debt Collection

Petmongkon Pornpichitsuwan, Krittin Pachtrachai and Touchapon Kraisingkorn

Voice agents built on large language models (LLMs) now handle customer calls, but in regulated applications completing the task is not enough: the agent must follow a conversational flow set by law. In debt collection, it must confirm it is speaking to the debtor before mentioning the debt, identify itself and the creditor, state the amount, and avoid prohibited language. Every creditor turns these rules into its own procedure and revises it. We ask whether a small, self-hosted policy can be trained once and then follow regulated procedures it has never seen. One machine-readable specification per deployment, FlowSpec, generates the policy instruction, gates the tools, supplies the supervision and defines the tests; the agent speaks only approved catalogue sentences. We fine-tune a 9-billion-parameter model on 2,506 conversations from 297 randomized procedures and evaluate a Thai outbound-call agent on four unseen deployment procedures (three debt collection, one non-debt), one from a wholly withheld lineage. Documented checks, fixed by the procedure, rise from 67–71 percents to 93–100 percents, while grounded checks, fixed by the conversation, already pass at 97–100 percents before training; scenario pass rate rises from 15/43 to 40/43. An audit of the shared specification finds 11 defects, 3 of which its own metric did not expose.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

Lotus: Learning to Understand Time-series by Learning to Forecast

Van Dai Do, Huu Hiep Nguyen, Minh Hoang Nguyen et al. (+2 more)

Time-series language models typically align numerical signals with language using labels, captions, or question-answer pairs. Although such supervision encourages linguistic representations, it does not explicitly require them to preserve information about how the underlying time-series signal evolves. We introduce forecast grounding, a pretraining principle that requires language-aligned time-series representations to remain predictive of the future evolution before they are adapted to downstream tasks. We instantiate this principle in LOTUS, which divides each time-series channel into non-overlapping patches and maps them into the LLM embedding space. It then predicts the next patch from the LLM hidden states of the time-series tokens, encouraging the learned representation to retain temporal information. To capture both temporal changes and periodic patterns, we also propose a temporal-spectral patch representation that combines each raw time-series patch with its spectral representation before mapping it into the LLM embedding space. Across three language model backbones and six downstream evaluation datasets, LOTUS achieves the best average performance with minimal trainable parameters, reaching 59.91% accuracy and outperforming the strongest matched baseline by 3.36%. Our representation analyses further show that LOTUS uses the numerical input and preserves temporal structure in its learned patch representations. These results establish forecast grounding as a promising pretraining principle, enabling compact time-series language models to outperform models 25 to 50 times their size on TSRBench.
Read the full abstract

All authors Van Dai Do, Huu Hiep Nguyen, Minh Hoang Nguyen, Manh Nguyen, Hung Le

Computation and Language

Fine-Grained Emotion Classification from Mobile App Reviews: An Empirical Study with Large Language Models

Quim Motger, Carlota Catot and Marc Oriol

Context: Fine-grained emotion classification of mobile app reviews enables requirements engineering activities that go beyond polarity-based opinion mining, including emotionally informed issue prioritisation and feature-oriented feedback analysis. However, automatic fine-grained emotion extraction from app reviews remains understudied. Objectives: Building on a previously published annotation framework and human-labelled ground truth adapted from Plutchik's taxonomy, this paper investigates how large language models can be leveraged for automatic multi-label emotion classification under severe class imbalance. Methods: We compare encoder-only fine-tuning under multi-label and binary-ensemble formulations, decoder-only zero- and few-shot prompting across open-source and proprietary models, and a catalogue of imbalance mitigation strategies (loss reweighting, resampling, generative data augmentation), with the synthetic-review generator and prompting strategy selected via an intrinsic augmentation-utility ranking. Results: Fine-tuned encoders trail the best decoder-only few-shot prompting (macro-F1 0.642) by a wide margin at baseline (multi-label: 0.387; binary ensemble: 0.450); pairing the best multi-label encoder with generative data augmentation and positive-weighted loss closes most of this gap (+0.204) at up to three orders of magnitude lower inference latency than the decoders, with the largest gains on the rarest emotions, from undetected to gains of up to +0.501 F1. Conclusion: Large language models make fine-grained, multi-label emotion classification of app reviews feasible for requirements engineering pipelines, with modest macro-F1, and the best formulation and mitigation strategy are backbone- and formulation-dependent. We release the experimental pipeline, synthetic corpora, and fine-tuned checkpoints for replication and reuse.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

How Causality Bridges the Semantic Gap

Shuhao Zhang, Xuran Zhou, Han Guo et al. (+2 more)

Numerical measurements capture how a system behaves, but often leave the meanings of its variables unspecified. Some variables are measured but never labeled, and others are never measured at all. Existing methods assign semantics to such variables by consulting general human knowledge, but this inherits its biases where that knowledge exists and offers nothing where it does not. We bridge this gap between measurements and their meanings with causal structure instead, reading a variable's semantics from how it acts on other variables. We formalize this as structure-constrained semantic alignment, in which the embedding of each unnamed variable is solved under the dependence relations implied by the causal graph, with the embeddings of a few known names as anchors. Accordingly, we build CausalBridge, a framework that discovers the causal graph from the measurements, latent variables included, solves for the embeddings under those relations, and expresses them as names through a language model. The causal structure reflects the mechanism that generated the measurements and is recovered from the measurements alone, which may make it the one source of information free of bias from human knowledge. We evaluate CausalBridge on five questionnaires and three robotics scenarios, with 20 to 90% of the variable names masked. It recovers the semantics of observed and latent variables more accurately than existing methods that rely on association, and its lead widens as less of the system is documented. The graph it discovers names variables as accurately as the documented one, and a new system is named in minutes and at a fraction of the cost of sampling methods. Once the semantic gap is bridged faithfully, machines can understand the world and take actions causally.
Read the full abstract

All authors Shuhao Zhang, Xuran Zhou, Han Guo, Pengtao Xie, Yujia Zheng

Computation and Language

Evaluating Multi-Dimensional Generalization of Large Language Models in Temporal Extraction Tasks

Fahmid Shahriar Iqbal, Ritam Dutt, Soumitra Das et al. (+2 more)

Time and event expression extraction are fundamental temporal reasoning tasks, but the problem remains difficult due to annotation ambiguity, domain sensitivity, and unstable model behavior. Existing evaluations focus on in-domain performance, offering limited insight into reliability under distribution shifts. We evaluate multiple model configurations across families, architectures, and reasoning strategies over four dimensions of generalization, examining transfer from base performance, cross-dimensional correlations, and the effects of scale, architecture, and prompting. This provides a systematic study of how prompted LLMs generalize in time and event expression extraction tasks. We find that strong base-task performance generally predicts better generalization. However, this relationship weakens under substantial distribution shifts. Inductive prompting performs most consistently across domain shift, adversarial perturbations, compositionality, and length increase, while gains from scale, architecture, and deductive and abductive prompting strategies are uneven and dimension-specific. We conclude that LLM generalization in temporal extraction tasks cannot be predicted from any single dimension alone and cannot be reliably inferred from in-domain or single-dimension evaluations, highlighting the need for reasoning strategies that generalize across dimensions.
Read the full abstract

All authors Fahmid Shahriar Iqbal, Ritam Dutt, Soumitra Das, Arnav Verma, Sagnik Ray Choudhury

Computation and Language

A generative-informed neuro-symbolic framework for syntactic ambiguity resolution: Evidence from Arabic DPs

Mohammed Damom, Mohammed Q. Shormani, Muneef Y. Alshawsh et al. (+5 more)

Syntactic ambiguity poses a persistent challenge for Arabic NLP, particularly in morphologically rich nominal constructions where multiple structu6ral interpretations may be compatible with the same surface sequence. This study proposes a generatively informed neuro-symbolic framework for resolving structural ambiguity in Modern Standard Arabic (MSA) DPs. The framework integrates generative syntactic notions with AraBERT by representing ambiguity as a candidate-based decision task in which linguistically motivated alternatives are explicitly constructed and evaluated through candidate-conditioned input representations. Findings indicate that the model achieved 96.88% accuracy, 95.92% macro-F1, 96.83% weighted F1, and 93.94% binary F1 on the unseen evaluation set. Class-level analysis revealed asymmetric performance, with recall of 99.71% for High/VP Attachment (N1) and 89.26% for Low/NP/Embedded Attachment (N2), indicating greater difficulty in recovering the embedded interpretation. The study concludes that formal syntactic representations can be operationalized within Transformer-based NLP as an explicit interface between linguistic structure and contextual neural modeling, providing a controlled and interpretable approach to Arabic syntactic ambiguity resolution and beyond.
Read the full abstract

All authors Mohammed Damom, Mohammed Q. Shormani, Muneef Y. Alshawsh, Ashraf A. Naji, Mustafa Ali Alhamzi, Fawwaz An-Nashef, Jameel Ahmed Elayah, Noman AL-Sayadi

Computation and Language · Cross-listed from cs.AI

Right Order, Wrong Scale: Auditing LLM Judges for Occupational AI Measurement

Harry Lyu and Neil Thompson

LLM judges are increasingly used to assess whether AI outputs meet workplace requirements, but agreement on response rankings does not establish agreement on acceptance rates or occupational aggregates. We introduce O*NET-BENCH, an audit suite derived from an existing survey of 45,796 worker ratings, and evaluate 33 pre-existing judge configurations across six model families on 4,501 test ratings. Twenty-five configurations achieve tie-aware pair accuracy of at least 0.60, although a train-fitted response-only TF-IDF baseline nearly matches the strongest judge. Despite this ordering agreement, judges estimate that 3.0%-97.9% of responses are acceptable, compared with 61.1% for occupation-matched workers. In one fine-tuned lineage, changing from pointwise scoring to a bundled few-shot/listwise protocol improves response ordering while reducing agreement with worker means at the task and occupation levels; this reversal replicates on a task- and worker-disjoint validation split under prespecified criteria. Cross-validated calibration largely removes mean bias, but calibrated scores explain at most 8.5% of individual worker-rating variance. Prediction-assisted estimation yields at most small precision gains at the studied label budgets. These results show that ranking agreement alone is insufficient for occupational measurement. Judges should be validated against the acceptance rates and aggregates their scores will be used to estimate.
Read the full abstract

Computation and Language

From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders

Pritam Deka

Typed decision models answer schema-constrained questions about a text in one forward pass and return probabilities meant to be thresholded. We ask whether biomedical sentence encoders trained for retrieval are good starting points for such models. We present SBERT2S1, which converts Sentence-Transformers encoders into bi-encoder, cross-head (C) and prior-fused residual (PFR) decision models, together with BIODECIDE, a biomedical typed-decision suite, and MEDLINE-S1, 243k training decisions derived from NLM indexing. Across six parent-retriever pairs, retrieval training improves zero-shot matching of content-bearing options. After fine-tuning, its effect depends on the head: across five pairs and three training-set sizes, retrieval training significantly helps PFR, which keeps the retrieval prior, in 10 of 15 comparisons, but helps C in one and hurts it in five. A matched grid of two heads and five training objectives shows that C outperforms PFR under every objective, and that the released RLCD recipe of open System One models trails cross-entropy by 2.5-3.0 points. The deficit stems mainly from its reward normalisation, which inflates the noisy score-function term 3.6-15-fold; an unbiased leave-one-out estimator recovers most of the gap. After temperature scaling, no objective is clearly better calibrated than cross-entropy. We release the code, the MEDLINE-S1 labels and a model.
Read the full abstract

Computation and Language

APDMem: Agent-Controlled Progressive Disclosure for Query-Adaptive Long-Term Memory

Chin-Lun Fu, Anagha Kulkarni, Hong Ni et al. (+1 more)

Personalized LLM assistants must recover sparse evidence from long conversation histories across queries of varying complexity. We introduce APDMem (Agent-controlled Progressive Disclosure Memory), a hierarchical long-term memory architecture that applies progressive disclosure to memory retrieval. Rather than relying on a flat memory store or fixed retrieval granularity, APDMem represents conversation history as four progressively detailed layers: thematic summaries, personalized key facts, turn-level evidence notes, and raw messages. At inference time, a controller applies progressive disclosure to the memory hierarchy: it first reads high-level summaries and drills into finer evidence only when needed. This creates an adaptive cost-fidelity trade-off: simple queries can terminate early, while complex temporal, multi-hop, or exact-evidence queries trigger deeper inspection. A note synthesizer converts retrieved evidence into a query-focused structure that consolidates facts, orders events, and flags contradictions before final answer generation. Experiments on LongMemEval show that APDMem achieves strong performance for long-context memory reasoning while accessing only 8% of the total conversations.
Read the full abstract

All authors Chin-Lun Fu, Anagha Kulkarni, Hong Ni, Behrouz Madahian

Computation and Language · Cross-listed from cs.LG

Capability Scaling-Down Laws for LLM Compression

Xueqi Cheng, Liang Wu, Kelly Wan et al. (+2 more)

LLM compression reduces inference costs and memory requirements, but selecting a method and configuration remains largely empirical because comparable resource reductions can produce different capability losses. We systematically investigate capability scaling-down laws for LLM compression across pruning, quantization, and distillation. Our framework measures capability loss in mathematics, code generation, and question answering, and relates these measurements to model size, training stage, compression settings, data availability, and training exposure. We develop simple predictive relations and evaluate their accuracy, measurement efficiency, and generalization to unseen configurations and model states. Sharing the density response across pruning levels halves the configuration measurements needed to fit a pruning predictor: on new Pythia states, on pre-registered OLMo-2 test states and under Wanda pruning, the compact relation matches a regression fitted with all measurements on math and code to within 0.020 nats per token, with coefficients refitted for each setting. Controlled distillation experiments show that the cost of heavy data reuse recurs across question-answering distributions, while the net benefit depends on the evaluation distribution. We further evaluate the decision value of these predictions by comparing numerical selection with configuration medians and fixed method priorities. Independent evaluations across two model families show that selection captures most of the available cross-method benefit for question answering within the tested candidate sets, where a fixed method priority attains the same regret, with smaller opportunities for mathematics and code. These results clarify the predictive scope of capability scaling-down laws and their use in compression method selection. Our code is publicly available at: https://github.com/LabRAI/scaling_down_law.
Read the full abstract

All authors Xueqi Cheng, Liang Wu, Kelly Wan, Liangjie Hong, Yushun Dong

Computation and Language

CUEing User Simulators: Calibrated User Embeddings for Multi-Turn Benchmarking

Anjali Kantharuban and Jonas Mueller

Recent benchmarks rely on user simulators to evaluate AI agents in multi-turn interaction. While existing simulation techniques demonstrate surface fidelity to human style and behavior, ecologically valid interactive benchmarking also requires alignment in when and how agents fail across simulated and real user populations. We find that existing simulators lack outcome calibration: agreement with observed success rates and failure patterns when real users interact with the same agent. We introduce Calibrated User Embeddings (CUE), a framework that both encodes observed sessions and samples continuous representations, then decodes them into persona commands to steer LLMs to act as user simulators without training. Through this, we evaluate user-conditioned replay of past sessions and aggregate metric agreement when sampling novel personas for the same tasks. On $\tau^2$-Bench, CUEd simulators commit fewer simulator-attributed errors and more faithfully reproduce real-user agent failure modes, aggregate success rates, and outcomes for specific task-user pairs than other persona-based simulation methods. These gains coexist with competitive user fidelity as measured using metrics established in prior work. After being fit to mostly customer support interactions, the same CUEd simulators generalize to document creation, math tutoring, and casual conversation, and remain effective across different simulator LLMs without CUE retraining.
Read the full abstract

Computation and Language

FinDialogLens: Event Extraction over Multi-Party Dialogue for Missed-Trade Identification in Financial Chatrooms

Chin-Lun Fu, Hong Ni and Behrouz Madahian

Multi-party financial chatrooms are vital for sales-and-trading professionals, but their complexity makes manual recovery of missed trades infeasible: each Request for Quote (RFQ) is an event whose final price and trade outcome appear many messages after the RFQ-trigger message (the inquiry message), interleaved with concurrent RFQs from other participants. We cast this as event extraction (EE) over multi-party dialogue and present FinDialogLens, a hybrid LLM pipeline in which compact fine-tuned classifiers act as inference-time scaffolds: they detect RFQ-triggers and price/trade outcome metadata, an RFQ-Level Module segments per-event RFQ windows, and a Trade Engine fills argument roles. With GPT-4o, FinDialogLens reaches 92.1% and 94.3% accuracy on final price and trade outcome, respectively, outperforming full-chatroom CoT prompting methods; fine-tuned open-source LLMs with as few as 3B parameters achieve comparable performance with modest in-domain data. To make the LLM-based solution practical at scale, a difficulty-aware router balances cost and accuracy by allocating RFQs between a low-cost rule-based engine and the higher-performing LLM-powered Trade Engine, cutting LLM calls by 85% on final price while recovering half of the accuracy gap to FinDialogLens (GPT-4o), saving over $300/day at our 70,000-RFQ/day scale.
Read the full abstract