artificialintelbulletin.com

Author: admin

  • Transformer-Based Masked Diffusion Models Provably Learn Distributions with Latent Structure

    Transformer-Based Masked Diffusion Models Provably Learn Distributions with Latent Structure

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Masked diffusion models (MDMs) generate discrete data by iteratively unmasking multiple tokens in parallel using a learned mask predictor. Their output carries two errors: parallelization error due to parallel sampling, and estimation error from the mask predictor learned with finite samples. Existing theory treats these errors separately, and the available statistical bounds typically scale with the number $V^L$ of length-$L$ sequences over a vocabulary of size $V$, suffering from the curse of dimensionality. In this work, we establish end-to-end statistical guarantees showing that transformer-based MDMs learn structured distributions with a sample complexity governed by their latent structure. Specifically, we model this structure through mixtures of hidden Markov trees, which capture latent dependencies within components and heterogeneity across components. When the data distribution is well approximated by such a mixture, a transformer-based MDM with parallel unmasking schedules achieves $varepsilon$-accurate sampling in KL divergence with $widetilde{O}(LK_{mathrm{par}}/varepsilon)$ training samples, where $K_{mathrm{par}}$ is polynomial in $L$, $V$, the number of hidden states, and the number of mixture components. The transformer network and sampling procedure require no knowledge of the tree topologies or component parameters, thereby adapting to the unknown latent structure automatically. Overall, our results provide theoretical justification for the empirical effectiveness of MDMs on real-world data.

    Submission history

    From: Jingda Wu [view email]
    [v1] Thu, 8 Oct 2026 17:54:49 UTC (90 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)

  • One Model to Loop Them All: Elastic Latent Recurrence for Language Models

    One Model to Loop Them All: Elastic Latent Recurrence for Language Models

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:A latent recurrent language model applies a shared block of layers repeatedly, so one checkpoint can trade inference compute for accuracy by changing the loop count. At one loop, however, such models typically fall well short of a non-recurrent model of the same size, a one-loop tax that makes a separate non-recurrent model the better choice whenever latency matters. We present Elastic Latent Recurrence (ELR), a recipe for training looped language models without this tax. At one loop, an ELR model costs the same to serve as a non-recurrent model of the same size and, trained on the same data, reaches similar accuracy. Each additional loop then buys more accuracy for more inference compute. Three choices make this work: a gated state update with a learned step size, loop counts mixed within each optimizer update, and conditioning each pass on its loop index and budget. The first two preserve one-loop accuracy, and the third lowers loss at every additional loop. Mixed depths can be introduced late in pretraining to limit the added training cost, and with zero-initialized conditioning added at the switch, this late schedule reaches lower loss than mixed-depth training from scratch. In MoE models, reusing first-loop expert routes on later loops cuts mixed-depth update time by 11.7% with no measurable change in accuracy. At inference, keeping only the last loop’s attention KV cache holds attention KV memory at its one-loop level while retaining most of the gain from additional loops. On 4B dense and 30B-A3B MoE hybrid models, ELR at one loop is 0.52 points below and 1.04 points above its matched baseline on a ten-task average, and four loops add 7.25 and 4.75 points. Carried through post-training, the recipe yields Koi-4B, a single checkpoint that scores an estimated 18 on the Artificial Analysis Intelligence Index at one loop, matching Qwen3.5-4B, and 29 at four loops.

    Submission history

    From: Shantanu Acharya [view email]
    [v1] Thu, 8 Oct 2026 18:05:42 UTC (296 KB)

  • Contextual Correction of Parallel Draftsfor Speculative Decoding

    Contextual Correction of Parallel Draftsfor Speculative Decoding

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Speculative decoding accelerates large language model inference by using a lightweight drafter to propose multiple tokens that are subsequently verified in parallel by the target model. Recent diffusion-based drafters make this paradigm particularly attractive: deep attention-based neural backbones provide strong predictions, while drafting multiple future positions in parallel avoids the sequential cost of autoregressive generation. This parallelism, however, comes at a cost. Tokens within a draft block are predicted without access to the realized identities of preceding draft tokens, removing intra-block dependencies that would naturally be available under autoregressive generation. We introduce DCOCO (DFlash Contextual Corrector), a lightweight mechanism that restores these dependencies while preserving parallel drafting. Rather than committing to each position’s top prediction, DCOCO retains multiple candidate tokens and evaluates their contextual compatibility across adjacent draft positions in parallel, enabling contextually informed token selection at low additional cost. Across four verifier backbones and ten benchmarks, DCOCO achieves the highest average throughput in every evaluated configuration, spanning greedy and stochastic decoding with and without extended reasoning, with speedups of up to 6.6x over vanilla decoding. Its advantage persists under concurrent serving: DCOCO outperforms the strongest competing method by 6-23% across 4-32 concurrent requests, while other sequential correctors fall below uncorrected DFlash at these loads. These results demonstrate that recovering intra-block dependencies through parallel contextual scoring delivers consistent acceleration across decoding regimes and serving loads.

    Submission history

    From: Roy Eisenstadt [view email]
    [v1] Thu, 8 Oct 2026 18:01:03 UTC (236 KB)

  • How AI Writing Assistance Distorts Perceptions of AI Users and Their Opinions

    How AI Writing Assistance Distorts Perceptions of AI Users and Their Opinions

    [Submitted on 24 Apr 2026 (v1), last revised 8 Oct 2026 (this version, v3)]

    View PDF HTML (experimental)

    Abstract:Hundreds of millions of people use AI for writing assistance. We asked how AI distorts perceptions of AI users and their opinions. Writers (N=2,939) wrote political opinion paragraphs with and without AI assistance. Separate readers (N=11,091) blindly evaluated these paragraphs and made inferences about the writers. With AI, writers seemed more opinionated, competent, positive, and demographically privileged. Writers objected to many of these distortions yet preferred AI-assisted text even when made aware of them. Reward models trained on our data mitigated objectionable distortions but reduced user acceptance. Additional readers (N=8,798) trusted AI-assisted writers more and were more persuaded by AI writing when AI was more distortive. Overall, distortions from AI writing assistance are pervasive, persistent under human oversight, and likely to have consequential effects on human behaviours and attitudes.

    Submission history

    From: Paul Röttger [view email]
    [v1] Fri, 24 Apr 2026 12:31:11 UTC (1,354 KB)
    [v2] Mon, 29 Jun 2026 11:51:27 UTC (1,436 KB)
    [v3] Thu, 8 Oct 2026 19:19:51 UTC (3,683 KB)

  • Smoothing the Top-k Exposure Boundary for Sparse Mixture-of-Experts

    Smoothing the Top-k Exposure Boundary for Sparse Mixture-of-Experts

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

    View PDF HTML (experimental)

    Abstract:Sparse Mixture-of-Experts models scale parameter capacity efficiently while maintaining a fixed compute budget per token. However, traditional training paradigms enforce a static choice of top-$k$ experts, which converts a continuous routing distribution into a rigid step function. This constraint introduces a brittle boundary where highly competitive experts are arbitrarily separated into full-supervision and zero-feedback zones based on minor score fluctuations. To address this issue, we propose Elastic Expert Routing, which stochastically samples the active expert budget from a localized discrete distribution centered at $k$. Over multiple training iterations, this mechanism softens the sharp threshold into a gradual probability distribution. Because the sampling neighborhood remains symmetric, this approach matches the expected computational cost of deterministic training, while preserving the inference budget. During supervised fine-tuning, elastic routing improves downstream macro-averages on OLMoE-1B-7B and Qwen3-30B-A3B by $+0.84$ and $+2.02$ points, respectively. In addition, in from-scratch pretraining, it reduces validation LM loss by up to $0.04$ and outperforms the static top-$k$ baseline by $1.6$ points on average across downstream tasks. Comparisons on Qwen3 fine-tuning and from-scratch pretraining further show higher average scores than Noisy top-$k$, expert FFN dropout, and DefaultMoE.

    Submission history

    From: Yunkai Chai [view email]
    [v1] Thu, 8 Oct 2026 09:31:19 UTC (157 KB)
    [v2] Fri, 9 Oct 2026 09:32:21 UTC (159 KB)

  • Unsupervised Multimodal Reward Modeling

    Unsupervised Multimodal Reward Modeling

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Vision-language reward models are trained on annotated preference pairs, which are expensive to collect, can disagree, and are ranked by benchmark accuracy. We build one with no human preference label: a linear head on an intermediate layer of a frozen Qwen3-VL backbone, trained to score a web caption above a perturbed version of the same caption; labeled data enters only to select the layer. The layer is the decisive choice, worth 5.8 points over the last layer, and one recipe transferred without re-tuning beats or at least matches GPT-4o on three benchmarks at 32B, 6.6 points ahead on VL-RewardBench, from a few thousand parameters at about a thousandth of fine-tuning compute. At an equal pair budget, perturbations teach a frozen head 3 to 5 points more than curated RLAIF-V pairs. Because the backbone is frozen, the objective is convex, and the perturbation sampler specifies the reward, which the benchmark does not identify: two heads trained on disjoint halves of the same perturbations tie on the benchmark and differ by 6.4 points in hallucination judging. Retraining the head on the policy’s own verifier-labeled candidates cuts caption hallucination by 5.1 CHAIR points and lifts image-generation pass rate by 5.4, while scoring lower on every reward benchmark.
    Together, an intermediate-layer probe, free same-image perturbations, and a verifier-driven bootstrap turn give a recipe for multimodal reward modeling that needs no annotator, costs a fraction of fine-tuning, and is specified by the comparisons it is trained on rather than by a benchmark rank.

    Submission history

    From: Yueying Li [view email]
    [v1] Thu, 8 Oct 2026 20:38:25 UTC (957 KB)

  • GeopoliticalQA: Benchmarking Retrieval and Generation over Intelligence Reports

    GeopoliticalQA: Benchmarking Retrieval and Generation over Intelligence Reports

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Geopolitical and strategic intelligence reports constitute a rich source of expert knowledge and have increasing value in the current political climate, but their heterogeneity and analytical complexity pose challenges for retrieval-augmented question answering. We introduce a benchmark for grounded question answering over 415 open-source geopolitical and strategic intelligence reports, comprising 5,525 document-grounded question-answer pairs spanning factual, inferential, comparative, and risk-oriented questions. We propose a document-aware generation pipeline that profiles reports before generating questions and validate dataset quality through expert annotation and an LLM-as-judge protocol. Using the benchmark, we evaluate 15 retrieval systems and six open-weight LLMs for RAG. We find that models perform similarly on document-level retrieval, while at the harder chunk level retrieval quality is not driven by embedding-model scale: bge-m3, a mid-size open encoder matches an order of magnitude larger ones and outperforms commercial embedding APIs. In contrast, answer generation remains challenging even given the gold evidence, particularly for inferential and risk-oriented questions.

    Submission history

    From: Veronika Solopova [view email]
    [v1] Thu, 8 Oct 2026 18:35:13 UTC (67 KB)

  • SynAct: Synergistic Active Feedback Learning for LLM Agents

    SynAct: Synergistic Active Feedback Learning for LLM Agents

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:LLM agents executing multi-step workflows fail in two fundamentally different ways: recoverable slips, which further attempts can repair, and knowledge gaps, which retrying cannot resolve. Expert feedback can address the latter, but expert attention is scarce. The key challenge is therefore not whether to ask for feedback, but where feedback can change the outcome. Existing approaches either avoid expert feedback, allowing agents to reinforce their own errors, or allocate feedback at the task level without identifying consequential decisions. We present SynAct, an active learning framework that treats expert attention as the binding constraint and allocates feedback at the level of individual decisions. SynAct ranks decisions by the mutual information between the selected action and the terminal outcome across repeated rollouts, queries an expert at the highest-value decisions, and distills the resulting corrections into transferable rules that the agent can reuse without expert access at deployment. We theoretically characterize what this signal identifies under finite repeated sampling: with k rollouts and at least m observations per action, outcome-constant decisions have zero mutual information, while outcome-separating decisions have mutual information at least H2(m/k) bits, yielding exact separation for any threshold t in (0,H2(m/k)). This also explains why success-based acquisition can be blind when all rollouts fail: success is constant even when failed runs reach different outcomes. Across four agent benchmarks and four models spanning three families, SynAct improves held-out accuracy in all 16 evaluated settings, by +11.7pp on average. The same signal transfers to inference-time routing, where SynAct outperforms budget-matched cascade, uncertainty, and expert-judge routers in 15 of 16 settings, while a single targeted correction outperforms seven additional unaimed retries.

    Submission history

    From: Ning Liu [view email]
    [v1] Thu, 8 Oct 2026 19:20:10 UTC (642 KB)

  • Kronecker Product Attention: Memory-Efficient KV Compression via Structured Factorization

    Kronecker Product Attention: Memory-Efficient KV Compression via Structured Factorization

    [Submitted on 8 Oct 2026]

    View PDF HTML (experimental)

    Abstract:Efficient autoregressive generation is increasingly limited by the memory footprint of the key-value (KV) cache, which grows linearly with sequence length, batch size, and model depth. Existing approaches such as Multi-Query Attention (MQA), Grouped-Query Attention (GQA), Multi-Head Latent Attention (MLA), and Tensor Product Attention (TPA) reduce this cost through different trade-offs, but often sacrifice expressivity, require additional positional-encoding machinery, or impose restrictive low-rank structure. We propose Kronecker Product Attention (KPA), which represents key and value activations as sums of Kronecker products of compact, token-dependent matrix factors. KPA strictly generalizes TPA: rank-one vector-outer-product factorization arises as a degenerate Kronecker split, while non-degenerate matrix factors can represent richer block-structured patterns. We further show that a structured rotary embedding operator matched to the Kronecker representation preserves the relative-position property of RoPE while rotating only the compact factors, removing the need for a decoupled RoPE branch as in MLA. Empirically, KPA matches or exceeds TPA, MHA, MQA, GQA, and MLA across four model scales and nine downstream benchmarks under identical training conditions, while achieving up to 10x KV cache compression. The Kronecker head split is a tuneable hyperparameter that allows practitioners to navigate the compression-accuracy trade-off at design time without architectural changes.

    Submission history

    From: Bahareh Nikpour [view email]
    [v1] Thu, 8 Oct 2026 19:23:56 UTC (227 KB)