Machine Learning
Direct Self-Evolving Optimization: Evolving LLMs without Challenger Training
Yuyang Deng, Yu Wang and Jiayun Wang
Self-evolving
language models improve by generating tasks and learning from their own feedback, but adapting the task generator often requires a separate challenger-training loop. Can we generate tasks adapted to the current solver without explicitly training a challenger? We introduce Direct Self-Evolving Optimization (DEO), which replaces challenger parameter updates with solver-guided task sampling. The KL-regularized challenger objective defines an exponential tilt of a fixed base task distribution. DEO uses this distribution as a sampling target: a frozen LLM generates and mutates tasks, the solver scores them, and an approximate Metropolis selection rule refines the training pool. Only the solver is trained. Theoretically, for an idealized variant that samples exactly from the tilted distribution, and under regularity, local gradient-dominance, and initialization conditions, we show that DEO learns distributionally robust
reasoning ability. In experiments, DEO achieves
reasoning performance competitive with R-Zero while using over $50\%$ less wall-clock training time, and improves
reasoning accuracy over a no-walk ablation. Replacing the task generator with a frozen API-only LLM further improves the local solver, illustrating a capability enabled by removing challenger training.
Read the full abstract ▾
Computer Vision
See, Measure, and Reason: Learning Visually Grounded Reasoning in Pathology
Chengyang Zhang, Wenchuan Zhang, Bo Li et al. (+8 more)
Pathological assessment relies on recognizing fine-grained visual details in histological images.
Vision-
language models (VLMs) increasingly support pathology interpretation, yet their ability to perceive these details remains inadequate. This weakness leads to inaccurate cellular observations that can persist even when final answers are correct. In this paper, we propose ASPECT to improve visually grounded
reasoning through explicit supervision of cellular appearance and abundance. ASPECT trains intermediate visual
tokens through pathology feature reconstruction,
cell feature
alignment, and count supervision. Three-stage supervised fine-tuning teaches the model to perceive, generate visual
tokens, and reason, followed by reinforcement learning that rewards answer correctness and consistency with reported measurements. We also introduce PathoVernier, a
benchmark of 759 expert-reviewed questions from five pathology
datasets covering four cellular composition tasks. It evaluates both final answers and intermediate measurements to expose errors hidden by answer accuracy. On PathoVernier, ASPECT achieves relative accuracy gains of approximately 19.2% over the strongest baseline, Gemini-3.1-Pro, and 99.3% over its Qwen3-VL-8B backbone, while reducing RAWR, which measures counting errors within correct responses, by 28.1% and 42.7%, respectively. ASPECT also improves over its backbone on three external pathology
benchmarks covering classification and question answering beyond cellular composition tasks.
Read the full abstract ▾
All authors Chengyang Zhang, Wenchuan Zhang, Bo Li, Mengran Li, Xinyu Liu, Jiaming Yang, Jie Chen, Zhang Zhang, Yuhao Yi, Hong Bu, Jiancheng Lv
Robotics
Xunyi Zhao, Jian Zhou, Sihao Lin et al. (+6 more)
Frontier models can now perform well on individual
embodied navigation tasks through multi-round
multimodal reasoning with simple tools. Across successive tasks, however, an
agent must also rely on an evolving map and earlier search records, both of which may be incomplete or conflict with new observations. We present NavHarness, a training-free
embodied harness towards lifelong
navigation that makes
memory processing part of the
navigation loop. During
navigation, its multi-round agentic session draws on maps, task records, and house knowledge, checking them against observations and recording corrections to guide its actions. NavHarness preserves this experience across fresh conversations for new tasks or recovery attempts, while outcome verification and run-end summaries support its later reuse. On GOAT-Bench, NavHarness improves s-SR over context-only independent sessions by 18.6 points with Astra and 22.6 with Opus 5. Using SLAM-estimated poses, NavHarness with GPT-6 Astra achieves state-of-the-art task success of 83.7 s-SR with 36.9 e-SR on GOAT-Bench and 85.9 s-SR on IR2R-CE. To understand these gains, we examine how experience is carried between sessions and find that structured recovery handovers outperform length-matched summaries. In extended deployments across houses, consolidation improves
navigation beyond retaining maps and task records, with case studies showing how
agents use earlier experience to interpret new goals, investigate unresolved questions, and resume failed searches. We suggest that progress towards lifelong
navigation depends on how successive
reasoning sessions build on prior experience, alongside improvements in single-task capability.
Read the full abstract ▾
All authors Xunyi Zhao, Jian Zhou, Sihao Lin, Gengze Zhou, Zerui Li, Xinyu Yan, Jiajun Liu, Anton van den Hengel, Qi Wu
Artificial Intelligence
BIABench: Evaluating AI agents on real-world bioimage analysis tasks
Zixuan Pan, Davide Panzeri, Lukas Johanns et al. (+5 more)
Artificial-intelligence (AI)
agents hold promise for automating bioimage analysis, yet no
benchmark evaluates whether they can carry out real-world analyses end to end. Such analyses are hard for
agents because 2D images, 3D volumes and time-lapse sequences are often too large to read as context, so an
agent must choose and run an analysis through code, specialized software and rendered views. Published studies make this capability testable, because each pairs raw images with a peer-reviewed result. We introduce BIABench, a
benchmark of 16 tasks reconstructed from published biological studies that retain their scientific questions, imaging data and ground truth. The tasks span eleven analysis subtasks and modalities from H&E histology to single-molecule localization microscopy. Each submission receives an outcome score, which compares the output files with the ground truth using field-standard metrics, and a process score, in which a
vision-
language model judges method choice and quality control against an expert-written rubric. We evaluated general-purpose and biology-specific
agents across several
language models, with repeated runs of every task. Routine two-dimensional tasks were solved well, but on some tasks that added a third dimension or a time axis no
agent scored above 0.19. Neither biological specialization, stronger models nor detailed expert instructions closed this gap. The
agents were also unreliable, with scores varying more between repeated runs of one
agent than between different
agents, and without ground truth a correct run could not be told from a wrong one by its process score or by the time spent. Released openly with its data and code, BIABench provides a verifiable framework for evaluating, and eventually training,
agents for reliable long-horizon bioimage analysis.
Read the full abstract ▾
All authors Zixuan Pan, Davide Panzeri, Lukas Johanns, Marilin Moor, Yu Zhou, Hedi Peterson, Yiyu Shi, Jianxu Chen
Artificial Intelligence
Query Expansion and Key Specialization in Transformer Attention Geometry
Vidit Gupta, Siddhesh Nadkarni, Mihik Chaudhari et al. (+2 more)
The projection of queries and keys are central to the
attention mechanism in
Transformer architectures. While they are mathematically symmetric, they play different roles in
attention mechanisms. The question of whether there is an effect from their functional distinction on their
geometric development in training remains unanswered. We investigate the problem through the training of small GPT-like
Transformers on character-level WikiText-103 for three different depths (4, 6, and 8 layers), three types of initialization for queries and keys, and four random seeds, resulting in 36 runs and 54 trajectories of average layers across seeds. We track the effective dimensionality of those layers using participation ratios and discover that effective dimension of queries expand while keys shrink, and that $PR_Q - PR_K$ is positive in all trajectories studied. In connection to
attention, the shrinking of keys leads to a narrower spectrum of $QK^\top$ and more peaked
attention weights. In order to determine if this connection is causal or coincidental, we directly control the spectrum of keys during training across five seeds: restricting it to make it shrink sharpens the
attention with high directional confidence, while keeping it constant to the level of initial dispersion makes
attention softer. Additional
token-level checkpoint analyses show that the monotonic paired-contrast trend is not universal across pretrained families, but survives as an early-training regime that later decays over a full pretraining run, and the link between interaction-rank geometry and
attention entropy remains visible in several models.
Read the full abstract ▾
All authors Vidit Gupta, Siddhesh Nadkarni, Mihik Chaudhari, Vinaya Sawant, Prachi Tawde
Machine Learning
Broken Symmetry in BF16 Attention: Why FlashAttention Gradients Blow Up Late in Training
Junlin Chen, Daize Dong, Huanwei Di et al. (+9 more)
BF16 is now standard in large-scale pretraining, including in fused
attention kernels such as FlashAttention, and these kernels are widely trusted. When we used FlashAttention-3 to pretrain a 450M-parameter
transformer on 50B
tokens, however, we ran into a problem: training was healthy for 25B
tokens, then the gradient norm grew a thousandfold and the loss ended 0.2 nats above FP32
attention, without a single NaN. Recomputing the
attention backward of just two layers in FP32 removes almost all of the excess gradient. Part of the cause is known: a fused multiply-add in the forward softmax, so far treated as an extreme-input NaN case and never fixed in FlashAttention-3. Repairing it stops the blow-up, but the query gradient is still wrong by more than its own size, and training still drives
attention logits to thousands of times their size under accurate gradients. The remaining error comes from a broken conservation law. The softmax score gradient sums to zero along every row, which makes the query gradient blind to where the keys sit as a group; rounding it to BF16 leaves a small nonzero sum that leaks the mean key into the gradient, and the leak grows exactly as late training makes keys large and
attention sharp. We introduce GProj (gauge projection), which restores the zero sum after the cast with two rank-one corrections per row. It cuts the remaining median query/key gradient errors from 219%/13% to 0.34%/0.37%, on par with FP32
attention, for 4.7% more time per training step. In matched from-scratch runs it trains to the same loss as FP32
attention, while FlashAttention-3 and key smoothing both destabilize.
Read the full abstract ▾
All authors Junlin Chen, Daize Dong, Huanwei Di, Haolong Jia, Jiawei Wu, Haotian Xie, Mingkai Zheng, Yang Li, Leshang Chen, Huishu Wang, Eric P. Xing, Hongyi Wang
Computer Vision
Scaling Versatile 3D Assets Editing with a Million-Scale Dataset
Badi Li, Tianxin Huang, Yu Zhou et al. (+3 more)
Although recent 3D generative models produce increasingly realistic assets, controllable 3D asset editing remains challenging. Existing methods are limited by scarce training data, insufficient source-aware modeling, and a lack of practical evaluation protocols. To address these limitations, we present Alchemy3D, a unified framework for training and evaluating versatile 3D asset editors that covers data construction, model architecture, and
benchmark evaluation. Specifically, we curate Alchemy3D-1M, a large-scale 3D editing
dataset containing 1.25M assets and 1.38M editing pairs across seven editing types. On this data, we train a family of generative flow models for general-purpose 3D asset editing. The model family supports image- and text-conditioned editing, few-step inference, and transfer to multi-view 3D part
segmentation. We further introduce GEdit3D-Bench, a large-scale, open-world
benchmark with a multi-dimensional evaluation protocol. Across existing and newly introduced
benchmarks, our method outperforms prior methods on most metrics of editing fidelity, source preservation, and visual quality.
Read the full abstract ▾
All authors Badi Li, Tianxin Huang, Yu Zhou, Wei-Shi Zheng, Yi Ma, Shenghua Gao
Robotics
Bayesian Active Learning for Intent Disambiguation in Interactive Robot Planning
Huao Li, Carson Sobolewski, Augustinos Saravanos et al. (+3 more)
Interactive
robot planning requires
robots to infer and execute human intentions from natural
language instructions that are often ambiguous, incomplete, or underspecified. Although large
language models (LLMs) provide a powerful interface for clarification, relying on the generative model to drive an multi-turn conversation can introduce systematic failures. We propose a
Bayesian framework that treats clarification as an active learning problem over grounded Signal Temporal Logic (STL) task specifications. Our method uses LLMs to initialize candidate formal specifications and translate informative contrasts into natural-
language clarification questions, while
Bayesian optimization maintains uncertainty estimation over user intent and selects queries that maximize information gain. After convergence, the inferred STL specification is passed to a formal planner to synthesize a verifiable
robot trajectory. Across four simulated and real-world task domains, our approach generally achieves higher task satisfaction and requires fewer clarification rounds than LLM baselines, while helping smaller models close the performance gap against larger
reasoning models.
Read the full abstract ▾
All authors Huao Li, Carson Sobolewski, Augustinos Saravanos, William Tan, John Karigiannis, Chuchu Fan
Data Structures and Algorithms
Rounding the Ball LP for Fair Max-Min Diversification
Julián Mestre, Lam Khai Trinh and Anthony Wirth
Given $n$ points in a metric space, partitioned into groups, $X_1,\dots,X_m$, and integer quotas, $k_1,\dots,k_m$, summing to $k$, the Fair Max-Min Diversification problem asks for a set of $k$ points, exactly $k_i$ from each group $X_i$, maximizing the minimum pairwise distance. Addanki et al. (ICDT 2022) described a ball LP for this problem and rounding
algorithms that yield a factor 2 approximation whose
fairness holds only in expectation and a factor 6 approximation with relaxed
fairness guarantees, as well as an $(m+1)$-approximation with exact
fairness. We introduce two new
algorithms. The first method refines the rounding of Addanki et al., yielding a 2-approximate solution that is $\varepsilon$-fair with high probability, meaning that from every group $X_i$, at least $(1-\varepsilon) k_i$ points are chosen. The second method in
polynomial time returns a 4-approximation to the optimal value of the exact
fairness version. In time $n^{O(1)} 2^{O(k)}$, which is fixed-parameter tractable in $k$, we achieve an exactly fair 4-approximation. Our method adapts the augmenting procedure behind Haxell's
theorem (Graphs Combin., 1995). This approximation factor does not depend on $m$. Moreover, we show that no rounding of the ball LP achieves a smaller factor with exact
fairness.
Read the full abstract ▾
Robotics
SAGE: Symbolic Action-Gating and Editing for LLM Task Planners
Trung Minh Bui, JongSul Moon, YoungOuk Kim et al. (+3 more)
Large
language models (LLMs) are now the default cognitive core of
embodied household
agents, yet the plans they emit are rarely checked against a grounded model of the environment before execution, and the task-success they report is often measured on
benchmarks so saturated that no method can be separated from another. We present SAGE (Symbolic Action-Gating and Editing), a single-LLM planner built from two lightweight mechanisms: a domain-agnostic symbolic gate (~250 lines of Python, zero
tokens, $O(|\pi|)$) that blocks precondition-violating actions with typed reasons as a runtime safety monitor, and a local edit that regenerates only the failed sub-goal's suffix, keeping completed and untouched work intact; a hybrid seed+live
memory store supports cold-start coverage. We evaluate under a leak-free protocol (leave-one-out
retrieval) over five open-weight models and a 75-task AI2-THOR
benchmark. On the standard
benchmark goal-completeness saturates (52% of instances trivially solved) and SAGE ties strong hierarchical baselines. On a harder, method-agnostic multi-goal composition, SAGE's completeness lead re-emerges large (+0.06 to +0.23 across four models). Under injected mid-execution failures, SAGE recovers as reliably as whole-plan replanners at 2.4-3.3x fewer LLM calls. As a verify-before-execute gate, the symbolic monitor blocks unsafe actions before actuation and raises simulator-reported step-success for every planner tested (up to +0.11), a signal the verifier never sees (non-circular). Because the gate calls no model (0.008 ms/plan), it is a safety layer that runs essentially free on the edge: SAGE planning reproduces its quality on a Jetson AGX Orin, where small-model verification helps most. We release the
benchmark, the leak-free protocol, the recovery and safety-gate harnesses, and a verifier-portability study (auto-induced on ALFWorld, 0.89 held-out).
Read the full abstract ▾
All authors Trung Minh Bui, JongSul Moon, YoungOuk Kim, Quang-Ngoc Phung, Se-Woong Jun, Dongin Shin
Machine Learning
ZeroCode: On-demand Error-Correcting Code Construction from the Zero Matrix via Reinforcement Learning
Ju-Hyeong Lee, Yongjune Kim, Sang-Hyo Kim et al. (+2 more)
Error-correcting codes (ECCs) are essential across diverse applications, from wireless communications and storage to
quantum computing, yet each application imposes distinct design requirements on the parity-check
matrix (PCM). To address these on-demand requirements in a unified framework, we propose ZeroCode, a reinforcement learning (RL)-based approach that constructs PCMs sequentially from the all-zero
matrix. ZeroCode formulates construction as a discrete sequential decision-making problem and uses proximal policy optimization with action masking to select valid edges. ZeroCode achieves a gain of approximately 1 dB over the prior RL-based construction method at a bit error rate (BER) of $10^{-4}$ for the (32,16) code and outperforms existing genetic, differentiable, and classical code-design methods in our experiments. Beyond optimizing decoding performance, the masking mechanism allows on-demand structural constraints, such as a maximum degree, 4-cycle-free structure, and quasi-cyclic structure, to be flexibly incorporated. Moreover, a single policy rollout yields a library of PCMs with varying edge counts, offering trade-offs between decoding performance and complexity without retraining. Overall, ZeroCode addresses diverse code-design requirements within a unified framework, providing solutions with optimized decoding performance under given constraints.
Read the full abstract ▾
All authors Ju-Hyeong Lee, Yongjune Kim, Sang-Hyo Kim, Dae-Young Yun, Hee-Youl Kwak
Machine Learning
Rotated Manifold Optimization for Low-Rank Adaptation
Yuhui Ding, Javier Zazo and James Hensman
We propose a novel optimizer for low-rank adaptation (LoRA) that explicitly incorporates the gauge
symmetry of low-rank factorization. Our optimizer extends recent
matrix optimizers for full-parameter training to the
manifold of fixed-rank
matrices by interpreting them as normalization under a rotated basis. We show how rotation and normalization can be integrated with the fixed-rank
manifold efficiently. Our optimizer converges faster to lower held-out loss and achieves better or comparable downstream performance on both supervised finetuning and reinforcement learning tasks.
Read the full abstract ▾
Hardware Architecture
Improving Indirect Branch Prediction in Interpreters via Hardware/Software Co-Design
Linfeng Zheng and Hiroshi Sasaki
Interpreters have a large indirect-branch footprint, requiring large predictor capacity for accurate prediction. We propose a hardware/software co-design in which a hardware lookahead engine, running ahead of the pipeline with software-provided bytecode metadata, supplies interpreter dispatch targets to the frontend. The engine requires only 1.3 KB of on-chip storage and changes to about 50 lines of CPython code. On 15 CPython server workloads, a 14 KB ITTAGE augmented with the engine reduces bytecode jump MPKI by 73.7% relative to a 16 KB ITTAGE baseline, yielding a 3.2% harmonic-mean IPC speedup.
Read the full abstract ▾
Artificial Intelligence
Maintaining Benchmarks Against Increasingly Capable Agents: Detection and Remediation of Unearned Passes
Weijun Luo, Kelvin Luu, Xinyi Liu et al. (+6 more)
Agentic
benchmarks guide model selection and training. Yet an
agent can pass a task without demonstrating the intended capability. Such outcomes constitute unearned passes; their proportion among all passes defines the integrity gap. As
agents improve,
benchmark surfaces that once seemed harmless can become exploitable, making
benchmark validity an ongoing maintenance problem. We introduce a process-verification framework that audits passing trajectories, distinguishes evidenced reward hacking from verifier weakness, and localizes exploitable surfaces for repair. Across 3,810 passing trajectories from 29 model-
benchmark cohorts, confirmed violations often increase with model generation but not monotonically. On SWEBench Pro V1.0, confirmed violation rates rise from 24% to 73% between Opus 4.7 and Fable 5 on matched tasks; later cohorts fall to 11% for Fable 5.1 and 0% for GPT-6 Astra. These comparisons are descriptive: configurations were not normalized, and the latest models also pass fewer exploitable tasks. Violations concentrate around a small set of recurring surfaces, especially unintended access to reference solutions through git history. Three repair case studies across two
benchmarks show why blocking a recorded exploit is insufficient: the same protected information can remain accessible through another route. Therefore, we combine minimal patches with exploit replay and fresh
agent evaluation,
auditing new passes under the original standard. No evaluated attempt against the final patches reached the protected channel, and every post-patch pass was judged legitimate.
Benchmark integrity requires ongoing maintenance: audit passing behavior, repair the enabling surface, and re-evaluate both exploit access and legitimate solvability.
Read the full abstract ▾
All authors Weijun Luo, Kelvin Luu, Xinyi Liu, Guangze Luo, Miguel Romero Calvo, Soham Dan, Daniel Yue Zhang, Ying Liu, Mohamed Elfeki
Robotics
RoboICL: Embodied In-Context Learning with GPT-6 Astra
Fangcheng Liu, Yeqing Shen, Anda Cheng et al. (+7 more)
General-purpose
vision-
language models offer a promising way to zero-shot
robot control: \gptastra{} excels at open-ended and
language- or image-conditioned manipulation but remains substantially weaker on high-precision and long-horizon tasks. We introduce RoboICL, an in-context
robot-control framework that narrows these gaps without
robot-specific parameter updates or a learned VLA. RoboICL separates demonstration context, which provides recorded examples when available, from interaction
memory, which accumulates the model's own actions and observed outcomes. Both use a shared observation–action–receipt–observation grammar. To preserve experience across task stages, RoboICL combines sampled demonstration blocks with bounded anchored
memory. Fixed anchors keep earlier rollout interactions available for in-context learning, while the latest interaction supports immediate error correction. Across 30 RoboDojo tasks, using zero shot for Open and one demonstration elsewhere, RoboICL improves on official zero-shot \gptastra{} by 20–27 progress-score points in every category. It leads the leaderboard baselines on
Memory and Open, achieves comparable performance to the strongest Precision baseline, and remains competitive on Long-Horizon. Its 30-task Overall score is 50.64, versus 33.68 for the strongest baseline. On a separate ten-task subset, RoboICL scores 60.60, within 2.00 points of the $\pi_{0.5}$ + \gptastra{} hybrid approach. On three real-
robot tasks, mean progress rises from 14.45 at zero shot to 63.33 at one shot and 78.89 at three shots. On two development tasks, optional Jev-gated action reuse reduces \gptastra{} calls by 33–48%. Code is available at \href{https://github.com/Mosi-AI/RoboICL}{https://github.com/Mosi-AI/RoboICL}.
Read the full abstract ▾
All authors Fangcheng Liu, Yeqing Shen, Anda Cheng, Weishi Mi, Chao Tang, Chenyuan Liu, Yushun Xiang, Tingguang Li, Yong-Lu Li, Yehui Tang
Artificial Intelligence
QuantaSpike: Short-Window Spike-Driven Quantization for Large Language Models
Bang Hu, Guowei Zhu, Changze Lv et al. (+4 more)
Large
language models (LLMs) achieve strong performance across many tasks but rely on dense multiply-accumulate (MAC) operations during inference, resulting in high energy cost. Spiking neural networks (SNNs) offer an event-driven alternative in which synaptic integration uses lightweight accumulation. However, spike-driven LLM inference remains difficult because outlier-heavy activations typically require long firing windows or auxiliary non-spiking paths. We propose QuantaSpike, a short-window spike-driven quantization framework for LLMs built around Logarithmic Ternary Integrate-and-Fire (LTIF)
neurons. LTIF uses ternary events with power-of-two membrane-response quanta, improving the information represented by each firing step while retaining shift-ACC-compatible computation. QuantaSpike combines this
neuron with group-adaptive gain and selective outlier admission: normal values use residual LTIF steps, whereas admitted outliers receive one additional onset spike before entering the same residual dynamics. Across OPT and Llama-2, QuantaSpike achieves state-of-the-art or competitive perplexity and zero-shot accuracy among spike-driven LLM quantization methods. It also transfers to newer dense LLMs, remaining close to the FP16 reference on Llama-3-8B and Qwen3-8B under the same four-step firing window. Analytical linear-energy projections show that QuantaSpike reduces the energy of one linear transformation by about $80.0\%$ on OPT models and $67.1\%$ on Llama-2 models relative to SpikeQuant, providing an accurate and energy-efficient spike-driven path for LLM inference.
Read the full abstract ▾
All authors Bang Hu, Guowei Zhu, Changze Lv, Xiaoqing Zheng, Fengzhe Zhang, Fan Zhang, Wei Cao
Computers and Society
Investigating Human–AI Discrepancies via Multiple-Solution Problems
Zihao Wang, Francesco Insulla and Andrea Montanari
Frontier artificial intelligence (AI) models are benchmarked on whether they reach a correct answer. Yet many problems admit several correct answers and repeated attempts, by different people or by the same model resampled, trace out a distribution over them. In this work, we ask whether human and model
reasoning lead to different distributions over valid solutions. Our testbed comprises 270
reasoning puzzles across five puzzle families. These multiple-solution puzzles each have 3 to 8 valid solutions and are simple enough that humans and models can solve them reliably. The resulting distributions differ markedly: models differ from one another, yet resemble each other far more than they resemble humans. Model distributions are, moreover, within every puzzle family, less diverse than human ones. We compare these discrepancies across puzzle categories, and trace how they respond to
reasoning-effort settings, to prompting, and to perturbations of the puzzle that leave its solutions unchanged. Together, these results point at significant differences between human and AI problem-solving processes, and their choice among equally defensible solutions. As progressive deployment of AI systems in society comes into focus, evaluating such differences (beyond one-dimensional accuracy metrics) is increasingly important. Data and code are available at https://hai-discrepancies.github.io/
Read the full abstract ▾
Computation and Language
Recursive LLM Degradation in Biomedical Question Answering: A Cross-Generation Study
Bibek Bhandari and Kshitij Lingthep
Repeatedly training
language models on their own generated data may create a synthetic-data feedback loop in which errors and distributional biases are reintroduced into subsequent training
datasets. This paper studies that process in biomedical question answering (QA) using PubMedQA and two Qwen2.5 model sizes, 0.5B and 3B parameters. The study compares a recursive synthetic-data condition, in which generation G(k+1) is trained on answers produced by G(k), against a Human-Control condition that repeatedly uses the original human training data. The study evaluates across four generations from G0-G3 with two random seeds (42 and 123) and a fixed evaluation set of 1,000 expert-labeled samples. The evaluation includes disease and chemical entity F1, context-supported rate, lexical and semantic similarity, answer length, repetition rate, and other evaluation metrics. The Recursive condition for both model sizes and both seeds showed larger declines than the Human-Control condition in disease entity F1, chemical entity F1, context-supported rate, ROUGE-L, and cosine similarity. Under the fixed no-repeat 3-gram decoding constraint, the main observed behavioral change was increased answer length, while the measured 3-gram repetition rate did not increase. The magnitude of the difference-in-change was larger for the 3B model than for the 0.5B model. This difference was particularly apparent in disease F1, context-supported rate, cosine similarity, and answer length. These results show domain-specific changes associated with using recursive synthetic-data training in biomedical QA, but do not establish clinical
hallucination rates or universal model collapse.
Read the full abstract ▾
Robotics
UMR: Universal Manipulation Representation
Song Liu, Linyi Li, Yanshun Zhao et al. (+11 more)
General-purpose
embodied manipulation hinges on a unified action representation that generalizes across embodiments and scales readily. Yet existing policies rely on embodiment-specific action spaces, making cross-embodiment demonstrations difficult to leverage at scale and limiting transfer to new embodiments and spatial variations. To this end, we introduce Universal Manipulation Representation (UMR), a unified action representation that enables zero-shot skill transfer from human demonstrations to heterogeneous
robots. UMR decomposes manipulation into two functionally distinct yet geometrically linked components: embodiment-agnostic World Flow, which describes task-relevant object motion in the world frame, and Ego Trajectory, which represents end-effector motion relative to the current pose. We instantiate UMR as World–Ego Point VLA (WEPVLA), a compact 0.5B-parameter policy that learns in the unified
geometric action space through a dual-stream Point Action Adapter and a unified Point Action Expert, with an $SE(3)$ conjugation coupling the two components. To improve data efficiency, we complement UMR with a Data-Efficient Strategy (DES) that diversifies object configurations through stage-aware point-cloud editing while preserving demonstrated contact geometry. In simulation, WEPVLA achieves average success rates of 97.5% on LIBERO and 85.7% on the 10-task RLBench
benchmark. In real-world experiments, a single policy trained on human demonstrations augmented by DES transfers zero-shot to diverse deployment conditions. With about 10 minutes of collected human demonstrations per task and no
robot demonstrations, it achieves 91.7% average success across six evaluation settings, compared with 60.8% for HumanEgo. Code and additional materials are available at https://umr-wepvla.github.io/.
Read the full abstract ▾
All authors Song Liu, Linyi Li, Yanshun Zhao, Rxuan Li, Xinrui Xu, Yi Ju, Yahui Deng, Senge Zhang, Guoyu Liu, Yixuan Li, Wuyang Zhang, Yao Li, Congcong Zhu, Jingrun Chen
Data Structures and Algorithms
Scheduling with Mandatory Breaks: NP-Hardness and an Additive-One Approximation
Shreyas Ghildiyal and Muralidhara V N
In classical fixed-interval scheduling, each job of a given set must be processed during a prescribed time interval. The goal is to assign each job to exactly one machine such that no two jobs assigned to the same machine overlap in their interiors, and the number of machines used is minimized. Without further constraints this is interval graph coloring and is solvable in
polynomial time through greedy approaches. We study a variant in which every used machine must remain idle during a contiguous break of prescribed length $x$ somewhere in the scheduling horizon. We show that this additional constraint makes the problem hard, except when $x\le1$. First, for every fixed $x\ge2$, deciding whether $k$ machines suffice for the assignment of a given set of jobs is NP-complete, even when all coordinates are bounded linearly in the number of jobs, so machine minimization is strongly NP-hard. Second, for $x=1$, we give a
polynomial-time
algorithm that solves the problem exactly. Finally, we give a deterministic
polynomial-time
algorithm that, on every feasible instance, outputs a schedule using at most $\OPT+1$ machines, where $\OPT$ is the true minimum. Unless $\mathrm{P}=\mathrm{NP}$, no
polynomial-time
algorithm guarantees $\OPT$ machines for any fixed $x\ge2$, so the additive guarantee of one is best possible.
Read the full abstract ▾
Machine Learning
Julian Boesch, Andrew Wee and Alexander Stranzl
Pretrained autoregressive
Transformers represent a large sunk investment in compute. Existing conversion methods reuse that investment by changing either the architecture (
attention to recurrence) or the objective (next-
token prediction to denoising), never both. We convert Qwen3 teachers at 1.7B and 8B into
attention-free, bidirectional, gated-delta-rule
diffusion students in three stages, so that each capability can be traced to the stage that kept or lost it.
Language modeling transfers only partially and in-distribution; in-context
retrieval does not transfer. On a multi-query recall probe where the teachers score 0.34-0.58, both converted students score 0.000, and
diffusion pretraining alone does not restore
retrieval. A
retrieval curriculum in the final stage, which gradually lengthens the gap between a key-value table and the queries that address it, restores it only stochastically: on a fixed schedule, one seed in three learns to retrieve. Advancing the gap only while a running accuracy estimate stays above a threshold works for all three of those seeds, holds on real text, and carries unchanged to 8B, where two of three seeds succeed. The third had not learned within its fixed 16k-step budget:
retrieval switches on abruptly at a seed-dependent step (6.5k and 11k in the other two), so a fixed budget can cut a late run off. One boundary survives every intervention: every model that learns
retrieval scores 0.000 on
tokens that never appeared in a
retrieval episode, and an arm that resamples the key and value
tokens every batch shows this is a coverage limit, not memorization of particular bindings. Separately, we convert a 7B code model into a 3:1 recurrent-
attention block-
diffusion hybrid over 85k steps and report two negative training results.
Read the full abstract ▾
Cryptography and Security
RADNPO: Reference-free Adaptive Negative Preference Optimization for LLM Unlearning
Shenghan Tan, Ziyi Zhou, Wenpeng Hu et al. (+1 more)
Large
language models (LLMs) can memorize sensitive, private, or copyrighted content during pre-training, making machine unlearning necessary for removing targeted knowledge. Recent preference optimization (PO)-based unlearning methods improve stability over gradient ascent (GA)-based methods by introducing
alignment-style objectives, which effectively suppress the probability of forget targets. However, target suppression alone does not sufficiently constrain the next-
token distribution after unlearning. Existing methods provide limited control over how suppressed probability mass is redistributed and insufficiently adapt forgetting strength to target confidence and distributional concentration. Even after target suppression, probability mass may remain concentrated on a few non-target
tokens, potentially producing repetitive or uninformative outputs. To address these limitations, we propose Reference-free ADaptive Negative Preference Optimization (RADNPO), which explicitly guides next-
token probability redistribution. Specifically, RADNPO contrasts each forget target with alternative
tokens favored by the current next-
token distribution and adaptively modulates
token-level forgetting strength using target confidence and next-
token concentration. Experiments on TOFU and MUSE demonstrate that RADNPO achieves a better trade-off between forgetting quality and model utility than current baselines.
Read the full abstract ▾
All authors Shenghan Tan, Ziyi Zhou, Wenpeng Hu, Mengyuan Zhang
Robotics
WAM-OPD: Sharpening World Action Models via On-Policy Distillation
Panjun Liu, Xiaohan Lei, Shiqi Zhang et al. (+8 more)
Pretrained world action models (WAMs) provide generalist capabilities across diverse robotic manipulation tasks, yet improving target-task performance to an expert level without degrading pretrained skills remains challenging. We explore on-policy
distillation (OPD) for WAMs and introduce WAM-OPD. WAM-OPD inherits the advantage of OPD methods that transfer task-specific teacher knowledge under the student's own induced distribution, rather than directly fitting the student to a narrow task-specific data distribution. However, in closed-loop manipulation, the observation histories change as the student policy evolves, requiring fresh environment rollouts to remain on-policy. Applying OPD to WAMs entails repeated data collection, which is costly even in simulation and often impractical on real
robots. To avoid repeated environment rollouts during
distillation, we introduce prefix-weighted trajectory replay (PWTR). PWTR uses a fixed trajectory pool composed primarily of initial-student rollouts, supplemented with task-specific teacher rollouts to broaden trajectory coverage. For each trajectory replayed from this pool, PWTR conditions the current policy on successive stored histories to generate fresh denoising paths, along which the task-specific teacher provides supervision. Although these denoising paths are refreshed as the policy evolves, the replayed environment trajectories remain fixed. PWTR therefore reweights per-decision
distillation losses using proxy importance weights derived from path scores accumulated over the trajectory prefix preceding each decision to mitigate the resulting shift in the history distribution. Simulated and real-world experiments demonstrate task adaptation without additional environment interaction during
distillation. In both settings, WAM-OPD improves target-task performance while retaining near-initial performance on tasks excluded from adaptation.
Read the full abstract ▾
All authors Panjun Liu, Xiaohan Lei, Shiqi Zhang, Yikun Wang, Yongxin Zhang, Mingyi Hu, Shida Sun, Jiateng Shou, Wengang Zhou, Jiajun Deng, Zhiwei Xiong
Artificial Intelligence
Evolving Support Priorities in Empathetic Reinforcement Learning
Pengyu Huang, Zhiyuan Han, Wenwen Tong et al. (+6 more)
We identify a fundamental mismatch in empathetic reinforcement learning: support priorities evolve with the dialogue state, yet existing methods typically optimize predefined reward specifications that remain fixed across turns. To model these evolving support priorities, we organize empathetic support along cognitive, affective, and proactive empathy, and propose Context-Adaptive Rubric
Evolution (CARE). At each turn, CARE generates a context-adaptive rubric by adjusting both the weights of these three empathy dimensions and their fine-grained evaluation criteria. The rubric generator is trained with turn-level rubric supervision and human preference data through supervised fine-tuning followed by preference-based reinforcement learning, and then serves as an adaptive reward interface for online empathetic RL. Integrated with both RLVER and MICA, CARE achieves state-of-the-art performance across SentientBench, EQBench3, and EMPA under three independent LLM judges. Notably, on EMPA, CARE improves EPM-Idx over the strongest baseline by at least 13 points under all three judges, including an increase from 28.11 to 83.54 under Gemini-2.5-Pro. Further analyses show that learned rubric priorities systematically vary across dialogue stages and user
emotions, demonstrating that CARE adapts what is rewarded as support needs evolve. Code is available at https://asa13opp.github.io/CARE-project-page.
Read the full abstract ▾
All authors Pengyu Huang, Zhiyuan Han, Wenwen Tong, Hewei Guo, Jiangnan Chen, Sirui Chen, Lewei Lu, Beier Zhu, Xun Yang
Computation and Language
SALMONN-duo: Adaptive Dual-System Coordination for Full-Duplex Voice Agents
Wenyi Yu, Siyin Wang, Terumi Chiba et al. (+5 more)
Full-duplex
speech large
language models (LLMs) enable low-latency, natural voice interaction. However, real-world
agents must also use tools and perform deliberative
reasoning-operations whose variable latency and computational cost conflict with the stringent timing requirements of real-time conversation. To reconcile these demands, we propose SALMONN-duo, an adaptive dual-system voice
agent inspired by dual-process theories of cognition. SALMONN-duo separates real-time interaction from deliberative computation by pairing an always-on, fast-thinking full-duplex
speech LLM (system 1) with a powerful asynchronous slow-thinking LLM
agent (system 2). Beyond handling real-time interaction, system 1 learns when to answer directly and when to delegate, remaining responsive during backend execution and seamlessly integrating returned information into the ongoing dialogue without exposing tool traces or losing conversational context. Evaluations on single-turn spoken question answering (QA) and multi-turn conversations demonstrate that adaptive delegation substantially improves accuracy on knowledge-intensive and multi-hop
reasoning questions, while knowledge-boundary-aware training avoids unnecessary system 2 invocations. On a customized version of $\tau$-Voice, SALMONN-duo further demonstrates its ability to complete environment-grounded, policy-constrained tasks through multi-turn interactions in realistic business scenarios. Finally, cost-aware reinforcement learning further enhances the trade-off between task performance and backend usage across the QA and conversation tasks, while improving task success and response safety on $\tau$-Voice with an acceptable increase in the delegation rate.
Read the full abstract ▾
All authors Wenyi Yu, Siyin Wang, Terumi Chiba, Xianzhao Chen, Xiaohai Tian, Jun Zhang, Lu Lu, Chao Zhang
Machine Learning
CasEm: A Cascade Architecture for Long-Horizon Neural Emulation
Zhaoyi Li, Jingtao Ding and Shihua Li
Autoregressive neural emulators can drift or diverge over long rollouts despite accurate short-term predictions. We introduce Cascaded Emulation (CasEm), a one-way rollout architecture that augments an existing full-state backbone with an independently evolving model of physically specified aggregates. Its forecasts guide corrections to full-state predictions, without feedback from the backbone to the aggregate model. Effective guidance requires aggregates that cover substantial backbone error, remain accurately predictable, and support useful full-state corrections. We derive a finite-horizon error bound that clarifies these three factors and use empirical diagnostics to guide subsystem selection. Across four ODE/PDE
benchmarks, CasEm reduces long-horizon rollout errors across diverse backbones and suppresses the trend toward error divergence in both
diffusion tasks using
Fourier neural operator backbones. In global
climate emulation, CasEm with a regional total-water subsystem reduces 10-year full-state time-mean error by 66.6% and 46.3% for frozen ACE and Spherical DYffusion backbones, respectively, while adding less than 3% to inference time.
Read the full abstract ▾
Cryptography and Security
Certified Multi-Source Integrity for Structured Agent Actions
Anmol Pandey, Aditya Jain, Liang Chen et al. (+2 more)
LLM
agents increasingly take privileged, often irreversible structured actions, such as paying an invoice. They assemble each action from action-critical fields in documents and tool outputs that an adversary can corrupt, and indirect
prompt injection can drive the model itself to extract attacker-chosen values. Current defenses gate on a source's trust label or certify free-text answer quality. None certifies the integrity of a coupled, policy-bound structured action under a corruption budget that accounts for shared upstream sources. We characterize when such an action is safely certifiable and give the maximally live safe certifier. It admits an action only when each field clears the rule its evidence structure supports: a bounded corruption radius over corruption-distinct evidence classes, counted by a minimum hitting set so that re-publishing or laundered copies cannot manufacture a quorum, deterministic reconciliation for complementary fields, and a trusted anchor where the evidence leaves a field single-sourced. We formalize two robustness notions, validate each mechanism by ablation, and measure how often the multi-source precondition holds on sanctions designations (70,966 entities) and software supply-chain provenance (450 packages). Under upper-bound proxies, genuine corroboration is a minority phenomenon in both, and naive attestation counting overstates it, since witnesses that look independent collapse to two corruption-distinct domains once shared origin is counted. Across five current models in a real
agent loop, a realistic injection fools every model but one and a naive
agent then executes the fraudulent action on most attacks. The certifier admits no unsafe action and recovers the correct value where corroboration permits, while action-gating and provenance baselines are broken in every world of our harness by some attack in its space.
Read the full abstract ▾
All authors Anmol Pandey, Aditya Jain, Liang Chen, Carsten Maple, Christo Panchev
Distributed Computing
Heddle: Learning Structural Templates for Parallelism Planning on Heterogeneous GPU Clusters
Taeyoon Kim, Yonguk Song, Seoyeong Choy et al. (+4 more)
Training large machine learning models on shared GPU infrastructures faces two challenges: (1) GPU availability shifts dynamically with varying resource demands from tenants, and (2) hardware heterogeneity accumulates as datacenters continuously adopt new GPU generations. Due to the vast search space induced by heterogeneous GPU types and node sizes, training planners must prune it aggressively to remain tractable, yet must also derive high-throughput plans promptly as cluster configurations change. Heddle achieves this goal through a learning-based planner that reduces the full planning problem to a search over pipeline structures. Heddle encapsulates planning decisions in a structural template and learns to construct plans from templates over diverse cluster configurations offline. This design is effective because structural decisions constitute the performancecritical core of a parallelism plan, while the rest follows by rule or from a small priced candidate set once the plan structure is fixed. Evaluation shows that Heddle matches or exceeds the best plan found by five existing planners across clusters with varying GPU types and node sizes for three models of different sizes by up to 84.5% in throughput on dense models and 4.6x on MoE models.
Read the full abstract ▾
All authors Taeyoon Kim, Yonguk Song, Seoyeong Choy, Hexiao Duan, Dong Li, Seo Jin Park, Myeongjae Jeon
Human-Computer Interaction
When Models Choose the Question: Pedagogical Constraints in Bottom-Up Multi-Agent Inquiry
Yeri Hong and Lauren Hyoseo Yoon
What shapes a model-generated inquiry when no discussion question is supplied? We introduce a bottom-up forum framework inspired by Philosophy for Children, in which
language-model
agents read a philosophical narrative, propose and select questions, and develop a shared conclusion without a privileged model facilitator or aggregator. Across 576 forums, contrasting Aristotelian value personas interacted with a blank-slate participant receiving no value-specific instruction. The blank slate remained neutral and was selected more often for conclusions than questions. Yet inquiry narrowed in both form and source: varied initial questions increasingly became either/or alternatives, while discussion concentrated on directions already explicit in the text. Our ECO framework traces this source focus by distinguishing explicit philosophical framing, characters' modeled inquiry, and open-ended narrative material. Across analyzed chapters, participants drew most often on explicit framing. We describe this as a pedagogical constraint: freedom to formulate questions did not necessarily produce freedom from source framing.
Read the full abstract ▾
Artificial Intelligence
Stashbird: Episode-Grounded Lifecycle Reconciliation for Conversational Agent Memory
Chidera Biringa, Lucas Yannul, Xiaowen Wang et al. (+5 more)
AI
agents require
memory that remains consistent as conversational evidence changes. We present Stashbird, an episode-grounded
memory system that uses provenance to reconcile derived graph and
retrieval state during updates and deletion. Repeated evidence extends support, corrections supersede prior active relations, and deletion removes unsupported state while preserving relations with remaining support. Under fixed semantic decisions, reconciliation produces correct final states in all 20 update sequences, compared with eight without reconciliation, and satisfies the defined checks across 263 deletion interventions. Across four long-term
memory benchmarks, Stashbird achieves competitive question-answering accuracy and strong model efficiency. On LoCoMo, it uses 76.4x fewer ingestion
prompt tokens than Graphiti and 8.1x fewer
retrieval prompt tokens than reproduced Hindsight
Read the full abstract ▾
All authors Chidera Biringa, Lucas Yannul, Xiaowen Wang, Marco Ayala, Nicholas Yi, Alex Moyse, Nishant Manchanda, Vivek Gupta