artificialintelbulletin.com

Tag: cs.CL

  • What Do Audio-Language Judges of Generated Music Listen For? Four Probes with Controls from Expert Ratings

    What Do Audio-Language Judges of Generated Music Listen For? Four Probes with Controls from Expert Ratings

    [Submitted on 9 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Open audio-language models are used as judges of generated music: given a clip and its text prompt, the model rates the clip, and the judge is validated by how well its ratings correlate with expert ratings. The correlation does not show what the judge responds to: the music, the text prompt, or the way the question is asked. We propose four probes that test each of these, three of them read against a control computed from the human ratings: removing the audio, replacing the clip with another model’s clip for the same prompt, swapping the order of two clips in pairwise judging, and rewording the instruction. We apply them to five open judges on all 2,748 clips of MusicEval, and the last two to the SongEval songs. Without audio, no judge exceeds what the prompt alone predicts from the human ratings, and for prompt alignment every judge falls below it. With audio, every judge follows the clip it hears when rating prompt alignment, two do when rating musical quality, and the strongest judge’s quality score tracks the experts’ alignment rating more than their quality rating. The pairwise verdicts of the four smaller judges follow the position of a clip in the pair (its slot) rather than its quality on both rating sets. For musical quality, rewording the instruction moves the correlation of most judges as much as changing the judge. MusicEval’s prompt assignment bounds any prompt-only judge at the model level. We recommend reporting the four probes with every music judge.

    Submission history

    From: Mateusz Modrzejewski [view email]
    [v1] Fri, 9 Oct 2026 08:56:54 UTC (50 KB)

  • Time Delay System for Sequence Modeling

    Time Delay System for Sequence Modeling

    [Submitted on 9 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Existing sequence models share a critical limitation: they lack an explicit, interpretable mechanism for content-aware temporal recall. LSTMs compress everything into a single hidden state, thereby losing information about specific past events. Transformers attend to all tokens indiscriminately at quadratic cost, while state-space models like Mamba compress history into a fixed-dimensional state without explicit content-addressable memory. To address this gap, we introduce a novel framework grounded in nonlinear time-delay systems, in which the effective delay is dynamically estimated based on the semantic similarity between current and historical inputs. Our approach features three key innovations: (i) a differentiable semantic lag estimator that computes content-aware soft attention weights over history, producing a weighted historical context and similarity score; (ii) a diagonal state space with closed-form cumulative sum solution, achieving O(1) per-step complexity with no sequential loops; and (iii) an interpretable fuzzy semantic modulation mechanism that learns context-dependent rules for balancing instantaneous processing versus historical recall, where rule centers correspond to prototypical semantic contexts. Experimental results on text classification and generation outperform state-of-the-art methods.

    Submission history

    From: Youshan Zhang [view email]
    [v1] Fri, 9 Oct 2026 01:43:37 UTC (583 KB)

  • Reflective Prompt Optimization of a Deep-Research Agent: A Topic-Disjoint, Multi-Seed Protocol for Measuring Transferable Gains and Prompt-Level Overfitting

    Reflective Prompt Optimization of a Deep-Research Agent: A Topic-Disjoint, Multi-Seed Protocol for Measuring Transferable Gains and Prompt-Level Overfitting

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Prompt optimization is an attractive alternative to reinforcement learning for adapting language-model agents, since it is cheap and produces human-readable artifacts. On factual tasks, however, it carries a failure mode that scalar-reward reinforcement learning does not. An optimizer that reads the training examples can write their answers into the prompt, so an in-distribution gain may be memorization rather than transferable behaviour. We apply reflective prompt evolution (GEPA) to the two system prompts of a multi-agent deep-research agent and evaluate it under a topic-disjoint split in which no topic is shared across train, validation, and test. Held-out citation reward rises by 0.079, from 0.702 to 0.781, with a bootstrap 95% confidence interval of [0.051, 0.107], and the gain replicates across three seeds (mean 0.107). Nearly half of the winning prompt is a memorized answer key, yet a controlled ablation shows this content earns no reward even on the exact training questions it encodes. The large validation-minus-held-out gap is therefore small-sample selection under a noisy judge, not memorization. Varying the verifier isolates what makes the gain transfer. Decomposed per-dimension feedback is the causal ingredient, and collapsing it to a single scalar reward makes held-out transfer negative. The gain persists under an out-of-family judge and is largely reproduced by a hand-written version of the discovered recipe. Credible evaluation of prompt optimizers therefore requires held-out topics, multiple seeds, and verifier ablations. We release the candidate prompts, lineage diffs, figures, tables, and analysis.

    Submission history

    From: Sahil Sen [view email]
    [v1] Thu, 8 Oct 2026 18:23:23 UTC (77 KB)

  • MADS: Multi-Agent Dialog Simulation for Robust Conversational Agents

    MADS: Multi-Agent Dialog Simulation for Robust Conversational Agents

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Task-oriented conversational agents remain fragile under realistic conditions, especially when users exhibit non-cooperative behavior. Existing function-calling benchmarks underrepresent adversarial conversation trajectories, thereby limiting the training resources available for developing robust agents. We present MADS (Multi-Agent Dialog Simulation), a synthetic data generation pipeline that produces structured, diverse dialogue trajectories. We inject adversarial conversation, producing validated conversation trajectories that capture realistic, unhappy paths. Using MADS, we construct MADS-Bench: a benchmark spanning six domains, 55 personas, and 1095 conversations, with approximately 66% of conversations containing failure-prone conversation scenarios. We evaluate the robustness of the trained models using a two-stage training approach that combines supervised fine-tuning (SFT) and reinforcement learning with verifiable rewards (RLVR).

    Submission history

    From: Arkajyoti Chakraborty [view email]
    [v1] Thu, 8 Oct 2026 22:08:10 UTC (420 KB)

  • COBALT: Contextual Bandit Learning with Offline Trajectories for Multi-Turn Code Generation

    COBALT: Contextual Bandit Learning with Offline Trajectories for Multi-Turn Code Generation

    [Submitted on 3 Feb 2026 (v1), last revised 8 Oct 2026 (this version, v2)]

    View PDF

    Abstract:Training large language models (LLMs) for multi-turn code generation typically relies on online reinforcement learning (RL), in which models learn from repeatedly generated online rollouts. Although effective, online RL can be costly and unstable. In this paper, we present contextual bandit learning with offline trajectories (COBALT), a simple method that reuses offline multi-turn trajectories to efficiently train LLMs for multi-turn code generation. Building on the observation that multi-turn code generation is a one-step recoverable Markov decision process, COBALT first collects code generation trajectories using a reference LLM and divides them into partial trajectories. Then, during online bandit learning, the LLM is trained to complete each partial trajectory through single-step code generation. On LiveCodeBench, COBALT substantially improves R1-Distill 8B and Qwen3 8B by up to 9.0 and 6.2 absolute Pass@1 points and matches the performance of two multi-turn online RL baselines using GRPO and VeRPO. We also perform a perturbation-based analysis of in-context reward hacking and show that Cobalt can improve robustness to misleading execution feedback. Overall, our results demonstrate that Cobalt is a promising solution for multi-turn code generation. Our code and data are available at this https URL.

    Submission history

    From: Ziru Chen [view email]
    [v1] Tue, 3 Feb 2026 18:08:41 UTC (265 KB)
    [v2] Thu, 8 Oct 2026 22:26:23 UTC (1,723 KB)

  • An Input-Dependent Readout Phase for Resonant Sequence Models

    An Input-Dependent Readout Phase for Resonant Sequence Models

    [Submitted on 26 Sep 2026]

    View PDF HTML (experimental)

    Abstract:Sequence models with a fixed-size state avoid the quadratic cost of attention, but face a design choice. An input-dependent state transition gives content-dependent memory at the cost of FFT-based training, while a fixed transition keeps the O(T log T) convolutional path but reads memory the same way for every input. ResonatorLM, a recent model of the second kind, reads a bank of damped complex oscillators through a fixed learned phase per head. We make that readout phase a function of the current token, which adds input dependence to the read only, keeps the convolution exact, and costs 0.2 percent more parameters. An analysis of the untrained bank shows that phase does encode elapsed time, but an ordinary token falls below the background within 2-8 steps and no choice of damping extends linear reconstruction beyond about 16 tokens. In 6M-parameter character-level models on WikiText-2, the input-dependent phase matches the original at the training length and removes its failure beyond it: trained on 256 tokens and tested on 1024, perplexity falls from 12.81 to 3.76. We trace the failure to long-memory heads that accumulate more history than training exposed them to. The learned phase tracks the current character rather than distance and does not improve retrieval.

    Submission history

    From: Shivesh Rane [view email]
    [v1] Sat, 26 Sep 2026 15:11:18 UTC (969 KB)

  • Does Financial Speech Need Continual Learning Regularization? A Study of EWC and mini-KFAC on Earnings Call ASR

    Does Financial Speech Need Continual Learning Regularization? A Study of EWC and mini-KFAC on Earnings Call ASR

    [Submitted on 9 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Earnings call transcription presents a realistic continual learning scenario where Automatic Speech Recognition (ASR) systems are updated with new recordings but risk catastrophic forgetting of previously learned financial speech patterns. Adapting Whisper-small from a general-domain speech corpus to the Earnings22 benchmark reveals the opposite: all evaluated methods exhibit improvement on held-out financial speech from the same domain during adaptation rather than forgetting. This task complementarity makes regularization counterproductive. Nevertheless, when regularization is warranted, a layer-wise mini-Kronecker-Factored Approximate Curvature (mini-KFAC) approximation seems to achieve lower word error rates (WERs) and more stable optimization than standard diagonal Elastic Weight Consolidation (EWC). The results suggest that task complementarity should be assessed before applying continual learning constraints, and prefer mini-KFAC EWC over diagonal EWC when needed.

    Submission history

    From: Homayoon Beigi [view email]
    [v1] Fri, 9 Oct 2026 11:34:13 UTC (1,103 KB)

  • Temporal Evidence Forests for Long-Horizon Multi-Party Conversational Memory

    Temporal Evidence Forests for Long-Horizon Multi-Party Conversational Memory

    [Submitted on 9 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Long-term memory has been widely adopted in LLM-based agents to support persistent interaction and personalized assistance. Long-term multi-party interactions, however, accumulate substantially more detailed information, with decisions repeatedly updated and states evolving across participants and interaction stages. Temporal structure is therefore crucial for tracing such memories: relevant information often appears densely within short intervals, while transitions among different states are anchored by a strict temporal order. Existing temporally aware memory frameworks primarily store time as metadata or use temporal signals to guide retrieval over preconstructed memory structures, without using temporal structure to both recover locally concentrated evidence and reconstruct its complete event progression. To address these limitations, we propose T-Mem, a temporal-structure-based memory framework that preserves interactions and organizes query-relevant evidence through two temporal mechanisms. Temporal Evidence Discovery exploits the local concentration of related messages to recover overlooked evidence, while Temporal Memory Structuring models state transitions and supporting relations over time and compresses low-information spans within the resulting forest. Evidence Forest Linearization then serializes the structured evidence for answer generation. Experiments on two long-term multi-party memory benchmarks demonstrate that T-Mem consistently outperforms competitive memory frameworks across LLMs, parameter settings, domains, and query types.

    Submission history

    From: Junyou Guo [view email]
    [v1] Fri, 9 Oct 2026 02:05:38 UTC (5,012 KB)

  • Measuring and Mitigating Solution Mode Collapse in RLVR

    Measuring and Mitigating Solution Mode Collapse in RLVR

    [Submitted on 8 Oct 2026 (v1), last revised 9 Oct 2026 (this version, v2)]

    View PDF HTML (experimental)

    Abstract:A language model (LM) can usually answer the same question in more than one way, but reinforcement learning with verifiable rewards (RLVR) is indifferent to which correct answer a model produces. A solution will earn the same reward whether it is the thousandth copy of a familiar answer or one the model has never produced before. Yet, there is potential value in having the model retain multiple correct solutions as it is trained. For instance, multiple modes may give users a choice and provide problem-solving strategies that improve overall model performance. Here, we introduce ModeBench, a benchmark of multi-solution tasks in which the verifier returns both correctness and mode discovered. We then use ModeBench to measure how solution diversity changes under RLVR post-training. We find that RLVR post-training concentrates probability onto fewer correct modes even as accuracy holds or improves, and moreover, that frontier models are already highly concentrated. We then introduce our solution, Re:Max, which stores one verified example per discovered mode in a replay buffer and trains on those stored modes uniformly. A solution found once is, therefore, practiced as often as one found repeatedly. Across three model scales, two RL objectives, and harder task constructions, replay improves both how often a policy succeeds and how many different ways it can succeed.

    Submission history

    From: Liv D’Aliberti [view email]
    [v1] Thu, 8 Oct 2026 01:25:16 UTC (744 KB)
    [v2] Fri, 9 Oct 2026 02:22:30 UTC (744 KB)

  • DistressBench: Evaluating Multi-Dimensional Crisis Support in Large Language Models

    DistressBench: Evaluating Multi-Dimensional Crisis Support in Large Language Models

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Language models have been adopted faster than evaluation practices that reliably assess their risks, and a growing share of use is affective rather than informational. Within that traffic users disclose suicide and self-harm (SSH) crises. Existing safety benchmarks score models primarily on refusal, but refusal is not the desired behavior here; the conversation requires a substantive supportive response. We introduce DistressBench, 718 English conversations with clinician-authored user turns spanning 23 SSH subcategories, each paired with a task-specific rubric of weighted atomic criteria across seven dimensions of crisis support and a clinician reference response. A model’s score is the weighted average over the criteria its reply satisfies, reading directly as the fraction of required care delivered. The best of 25 frontier models satisfies 88.4% of weighted criteria and the median model 71.4%, against 98.7% for clinician references scored on the same criteria. Pooled across all 25 models, in 35.3% of conversations where a model satisfied every criterion for recognizing the crisis, it failed to route the user toward human help. The shortfall concentrates in the dimensions requiring the model to act rather than withhold. Contrasting each dimension’s median model against its best separates capability limits from choices: de-escalation reaches 74.6% in the strongest model against a median of 46.0%, and safety disclaimers 84.6% against 8.3%. Multi-turn conversations score below single-turn ones (68.5% against 74.2%). We release the corpus, rubrics, judge calibration labels, and a companion annex of 20 long-horizon red-team transcripts. These results argue for evaluating crisis safety on the care a model delivers rather than on what it declines to say, and for treating de-escalation and resource routing as deployment requirements.

    Submission history

    From: Drew Rein [view email]
    [v1] Thu, 8 Oct 2026 22:09:12 UTC (328 KB)