Computation and Language
QuantCode Model: Specializing Language Models for Executable Algorithmic Trading Code
Alexey Chernysh, Orkhan Ekhtibarov and Dmitry Zmitrovich
Large
language models are strong general-purpose code generators, but executable algorithmic trading remains a demanding specialization target: a model must translate a natural-
language strategy specification into correct program logic for a specialized trading framework, execute on historical data, produce trades, and remain semantically faithful to the request. We study two complementary mechanisms for specializing
language models for this setting: continued pretraining on algorithmic-trading framework code and supervised fine-tuning (SFT) on
agent-validated request-to-code pairs. Evaluation is centered on QuantCode-Bench, our 400-task
benchmark for Backtrader strategy generation, together with a repository-level SWE-bench-like track. Continued pretraining improves single-turn Judge Pass from 41.5% to 47.5% for Qwen3.5-397B-A17B and from 27.8% to 33.0% for Qwen3.6-35B-A3B. SFT applied after continued pretraining yields a larger gain for Qwen3.6-35B-A3B, reaching 58.2% Judge Pass and 83.5% successful backtests; in agentic evaluation it raises first-turn success from 22.3% to 58.3% and final success after up to 10 turns from 47.5% to 79.5%. Continued pretraining alone improves first-turn agentic success but lowers final success after repair from 47.5% to 32.5%, consistent with degraded instruction following, whereas SFT improves both. We also identify a capability-retention failure: domain specialization degrades parser-conformant structured tool calling, and targeted recovery SFT restores tool-call formatting but not the base checkpoint's repository-level
agent performance. The results show that framework-oriented pretraining, validated SFT, and explicit capability-retention evaluation address distinct failure modes in domain-specific executable code generation.
Read the full abstract ▾
Robotics
Communication-Free Distributed Multi-Robot Task Allocation under Partial Observations Using Labeled Multi-Bernoulli Filtering
Takumi Ito and Akiya Kamimura
This paper proposes a communication-free multi-
robot task allocation framework based solely on local observations. In this study, tasks are defined as reaching target locations. Each
robot estimates the positions of neighboring
robots using a Labeled Multi-Bernoulli (LMB) filter and independently assigns tasks through a greedy auction-based strategy. By continuously updating state estimates and reallocating tasks during execution, the proposed method enables decentralized coordination without explicit communication. Monte Carlo simulations demonstrate that the proposed method enables effective cooperative task allocation without inter-
robot communication while remaining robust to measurement clutter and observation uncertainty.
Read the full abstract ▾
Cryptography and Security
SEW: Style-Encoded Watermarking of LLM-Generated Code
Soohan Lim, Hyundong Jin and Yo-Sub Han
Code watermarking supports provenance tracking for code generated by LLMs. Modifying
token selection to embed watermarks as an LLM generates code can create a trade-off between detectability and functional correctness. Other methods instead watermark completed code using predefined transformations or trained neural models. Recurring patterns can make watermark choices predictable across programs, while treating patterns common in unwatermarked code as watermark evidence can cause false detections. We therefore introduce SEW, which embeds and detects watermarks in already generated code through three components: (i) code style rules collected from style guides and transformation rules, with style choices determined by a secret key and each program's structural context; (ii) style-preference calibration, which evaluates watermark evidence using style probabilities estimated from human-written code; and (iii) context-aware style aggregation, which combines evidence from structurally matching locations assigned the same code style choice, preventing repeated applications of that choice from inflating watermark evidence. On CodeContests across three LLMs and three programming
languages, SEW achieves a mean relative improvement of 12.44% in TPR@FPR5% over the baselines and is robust to four non-LLM code-editing attacks, with only a 0.94% mean relative decrease. Our code is available at https://github.com/suhanmen/SEW.
Read the full abstract ▾
Human-Computer Interaction
Aligning the Incomplete: Joint Distribution Calibration for Multimodal EEG-Eye Emotion Recognition
Yang Wu and Jinpeng Li
The success of cross-subject
multimodal emotion recognition hinges on maintaining the consistency of the joint data distribution across individuals. However, real-world deployment frequently triggers the asymmetric joint distribution collapse:
EEG signals suffer from severe cross-subject distribution shifts, while eye movements sensors are susceptible to packet loss and tracking failures. Existing methods treat domain adaptation and missing-modality imputation as disjoint tasks. Consequently, they fail to resolve the compounded errors when both degradations co-occur, either propagating domain shifts through imputed signals or destroying the joint decision boundary. To tackle this unified challenge, we propose GUARD (Gradient-guided Unsupervised Asymmetric Recovery of Distributions). First, GUARD establishes a reliable anchor
manifold in the source domain by employing a theoretically grounded gradient-weighted objective, which forces the robust
EEG modality to preemptively entangle task-discriminative ocular features. Next, to structurally recover the collapsed joint distribution, we constrain a generative module with downstream perceptual losses, prioritizing
emotion-discriminative semantics over mere signal fidelity. Finally, we formulate target-domain adaptation as an ill-posed inverse problem. By driving a cycle-consistent flow, we achieve unsupervised calibration of the recovered joint distribution directly on the target-domain
manifold. Extensive experiments demonstrate that GUARD significantly outperforms state-of-the-art methods, maintaining resilient discriminative performance even under complete auxiliary modality failure. Our code and models are made publicly available to ensure complete reproducibility.
Read the full abstract ▾
Databases
Recommendation Systems for Exploratory Data Tasks
Anna Fariha
A large class of data-centric tasks is exploratory, where users iteratively steer workflows, refining subjective goals as new insights emerge. These Exploratory Data Tasks (EDTs) are performed by millions of users with varying levels of expertise to understand unfamiliar data, discover trends, and identify evidence that informs critical decision-making. However, a key challenge in EDTs is the enormous space of possible actions that one can take at each step: users struggle to choose among thousands of joins, transformations, and aggregations, causing "exploration paralysis". Because EDT workflows are interconnected, each choice impacts subsequent exploration, and suboptimal choices can lead to inefficiency, missed insights, confirmation bias, and incomplete coverage. This calls for intelligent recommendations that efficiently guide users toward optimal EDT actions. We envision recommendation as a core capability of data systems, proactively guiding users toward promising actions and thereby lowering the barrier to exploratory data tasks. EDT recommendation is challenging because the action space is combinatorial and actions are data-dependent, which require costly materialization. Moreover, recommendation often involves bundles or sequences of actions across interdependent tasks, requiring coordination across tasks. In this paper, we present our
vision of EDT recommendation systems along two axes: single-task vs. multi-task settings and single-action vs. multi-action recommendations. We outline a research agenda that progresses from recommending individual EDT actions to constrained bundles and sequences of actions, and ultimately to coordinated recommendations across interconnected EDTs. We identify research directions for incorporating various contexts (user, data, task, and ecosystem), addressing efficiency challenges, and coordinating across tasks.
Read the full abstract ▾
Machine Learning
Awakening of the Buddha: Subspace Learning During Population-Loss Plateaus
Akash Kumar
Population loss can remain nearly constant while a neural network learns a substantially more predictive representation. We establish this separation for two-layer ReLU and leaky-ReLU networks trained on Gaussian inputs by simultaneous fixed-step population gradient descent on all parameters. For structured additive teachers whose links are positive mixtures of Gaussian-damped cubics in $H^1(\gamma)$, we give explicit conditions under which small IID Gaussian initialization yields a high-probability guarantee: at a checkpoint during a high-loss plateau, minimum
alignment between the rank-$r$ teacher subspace and the leading $r$-dimensional eigenspace of the predictor's average gradient outer product (AGOP) increases by at least $1/2$, and the minimum refit MSE under unchanged coefficient budgets decreases by more than $0.399$, both relative to initialization. The same trajectory subsequently attains a trained loss below every value in the plateau window. A complementary result treats unequal-weight cubic teachers and small additive Sobolev perturbations using projected-feature refits. For SwiGLU networks with an exactly fitted intercept, we prove leading-AGOP
alignment during a loss plateau at fixed width and dimension as Gaussian initialization vanishes, for square-integrable teachers with nonzero Hermite content of degree one, two, or three. A rank-one cubic specialization also gives simultaneous unrestricted-refit gains at a prescribed width. An approximation lower bound further shows that certain interaction targets retain nonzero error when ridge
neurons are restricted to shared orthogonal axes within the teacher subspace. Population-moment experiments with ReLU students across 21 teachers and 50 initializations per teacher complement the analysis.
Read the full abstract ▾
Artificial Intelligence
Inferring Causal Relations between Two Sequences of Events with Language Models
Nishchal Prasad, Eric Gaussier, Emilie Devijver et al. (+3 more)
Causal AI is a branch of Artificial Intelligence which helps understand and reason about cause and effect relationships, not just patterns or correlations. Causal discovery aims to infer elements of the underlying causal structure–often represented as a directed graph–from observational and, when available, interventional data. While causal discovery is the fundamental step for moving beyond mere associations toward genuine understanding, and thus the basic building block of causal AI, it becomes intrinsically difficult when causal relations must be inferred from single observations. In such situations, standard causal discovery methods cannot be used and one has to identify causal relations from limited amount of information. This is typically the case for, e.g., sequences of events produced by different alarms which need to be analyzed on the fly to detect abnormal phenomena, which are usually rare. We show in this study that it is possible to leverage the predictive power of Large
Language Models (LLMs) to infer causal relations between only two sequences of events. This approach, which is validated on both synthetic and real data, provides better results than standard causal discovery
algorithms on several time series data, even though these data were converted into smaller, single observed sequences.
Read the full abstract ▾
All authors Nishchal Prasad, Eric Gaussier, Emilie Devijver, Alexander Obeid Guzman, Armen Aghasaryan, Gregor Gössler
Machine Learning
No Task Vector Is an Island: A Comprehensive Study on the Composability of Task Vectors from On-Policy Distillation
Jingang Zhou, Feiyu Han, Han Zhu et al. (+6 more)
Task vectors provide a simple mechanism for composing learned capabilities through model merging. However, the composability of task vectors produced by on-policy
distillation (OPD) remains largely unexplored. OPD trains a student using teacher feedback on student-generated trajectories, yielding parameter updates that differ from those produced by the teacher model, usually by reinforcement learning (RL). We therefore ask whether OPD task vectors can complement their RL teacher updates and compose effectively across tasks. Across five domains and two model architectures, we find evidence for both forms of composability. Within a task, merging OPD and RL task vectors can outperform both constituent models, even when the OPD student is weaker than its RL teacher. Across tasks, OPD task-vector compositions achieve higher average scores than corresponding RL compositions in seven of eight backbone-merging-rule comparisons. Parameter-space analyses reveal substantial non-collinearity between OPD and RL updates. Experiment in CODE domain on SMOLLM3-3B shows that the combined direction outperforms either constituent direction at the tested global update norm, supporting directional complementarity in this configuration. Across tasks, OPD updates also show lower overlap among the top-10% feed-forward channels ranked by update energy. Together, these results show that weaker standalone performance does not imply weaker task-vector composability. OPD task vectors can complement stronger RL teacher updates and combine effectively across tasks, highlighting composability as a distinct property for understanding and evaluating post-training updates.
Read the full abstract ▾
All authors Jingang Zhou, Feiyu Han, Han Zhu, Yuyi Zhou, Ruiyang Zhang, Jian Xu, Sirui Gao, Qingpei Guo, Xu-Yao Zhang
Robotics
IronMind: Scaling Humanoid Dexterous Manipulation via Camera-Space Ego-Centric Pretraining
Huimin Pan, Yufan Ren, Kunpeng Song et al. (+16 more)
Egocentric human
video offers a scalable data source for dexterous manipulation, yet using it to train
humanoid robots presents two challenges: (1) an embodiment gap, as human hands differ structurally from
robot end-effectors and low-cost
egocentric recordings lack the torso kinematics required by conventional retargeting; and (2) heterogeneous data quality, including noisy hand-pose tracking and weakly aligned text annotations. We introduce IronMind, a
vision-
language-action (VLA) model that uses
egocentric human
video and heterogeneous
robot data to pretrain policies for
humanoid dexterous manipulation. To bridge the embodiment gap, IronMind bypasses explicit body-retargeting by using a camera-space action representation, the native reference space of
egocentric video, and semantically aligning
robot and human action dimensions. Across total pretraining budgets from 250 to 10,000 hours, validation loss decreases approximately log-linearly with data scale. Larger pretraining budgets also improve out-of-distribution real-
robot manipulation after post-training: across six challenging tasks with unseen objects, affordances, and
reasoning prompts, the 10,000-hour model achieves a 55.0% success rate, compared with at most 11.7% for every pretraining budget up to 5,000 hours and 5.0% without pretraining. At the same pretraining budget, the camera-space action representation also outperforms the torso-frame baseline. Together, these findings support pretraining with a camera-space action representation on large-scale human
egocentric data as a scalable foundation for
humanoid robot manipulation.
Read the full abstract ▾
All authors Huimin Pan, Yufan Ren, Kunpeng Song, Siyang Wang, Xiwen Zhang, Xiaoyun Hu, Zhuoxu Duan, Hanrui Zheng, Jialeng Ni, Nathan Zhao, Sibo Ma, Zhenxuan Fan, Zhongyang Che, Danny Bao, Jiacheng Wei, Jerry Bai, Xiaoyu Yue, Xiaoyang Guo, Chenyi Chen
Artificial Intelligence
Advancing Entropy-Level Credit Assignment in RLVR via Proximal Entropy Policy Optimization
Yun Kim and Nojun Kwak
Value-model-free RLVR methods such as GRPO assign uniform advantages to all
tokens in a rollout, ignoring that
tokens contribute unequally. Recent methods use
token entropy as an importance proxy but compute it globally across the batch, conflating importance with
prompt difficulty and positional trends. We argue that importance should instead be measured relative to the local context of each
token. We introduce proximal
entropy, a local measure of
token importance relative to neighboring
tokens, and prove it is invariant to both confounders. Proximal
Entropy Policy Optimization (PEPO) uses it to weight per-
token advantages and outperforms GRPO and
entropy-based baselines on mathematical
reasoning across Qwen3-1.7B, Qwen3-4B, and Llama-3.2-3B-Instruct. We also show the formulation generalizes to other
algorithms where substituting proximal
entropy into existing methods improves, and applying it to single-stream RL succeeds where global
entropy fails.
Read the full abstract ▾
Systems and Control
Sharing the Gains of Aggregation: Cooperative Imbalance Cost Allocation
Asmus Winther Eriksen and Jalal Kazempour
A facilitator is an intermediary that offers renewable producers and consumers fixed-price contracts and, acting as their balance responsible party, manages the residual imbalances in the
market. Pooling imperfectly correlated residuals nets consumers' imbalances and reduces the portfolio's total imbalance cost, but raises an allocation question: how should these savings be divided among heterogeneous consumers? We formulate this as a cooperative game, the imbalance netting game, and compare six allocation mechanisms under a two-price imbalance settlement in terms of computational requirements, budget balance, group rationality, and additivity. We establish three analytical results: a necessary and sufficient condition for budget balance of the marginal cost contribution mechanism, group rationality of the marginal cost contribution and Vickrey-Clarke-Groves mechanisms, and a necessary and sufficient condition for when the Gately point is well-defined. On Danish 2024 data for 19 consumers, aggregation reduces imbalance cost by 10%, but how these savings are shared depends strongly on the allocation rule. A flat-rate allocation proportional to consumption, used as a
benchmark for socialized imbalance pricing, charges some consumers more than twice their standalone cost, whereas all the game-theoretic mechanisms studied produce stable allocations in practice.
Read the full abstract ▾
Machine Learning · Cross-listed from stat.ML
Towards Optimal Inventory Control under Censored Demand: A Biased Sample-Average Approximation Approach
Yuxuan Han, Xiaoyu Fan, Jiawei Zhang et al. (+1 more)
We study data-driven multi-period lost-sales inventory control under censored demand, where a stockout reveals only that demand exceeded the stocking level. We develop a unified, model-based framework for policy learning from censored data, built on a new cost decomposition for base-stock policies and a biased sample-average approximation (SAA) approach. The cost decomposition allows us to propose a new coverage condition under which censored observations are informative enough for sample-efficient policy learning. Guided by this coverage condition, we design two biased SAA
algorithms: an upper-biased one that achieves near-optimal sample complexity under the offline coverage condition, and a lower-biased one that actively generates the required coverage and achieves near-optimal regret online. More broadly, this biased SAA approach provides a general principle for implementing pessimism and optimism under censored feedback, which may be of independent interest.
Read the full abstract ▾
All authors Yuxuan Han, Xiaoyu Fan, Jiawei Zhang, Zhengyuan Zhou
Systems and Control
Learning to Fix: Optimisation-Aware Machine Learning for Accelerated Unit Commitment
Benjamin Fritz, Andreas Makrides, Maryam Fetanat et al. (+1 more)
Unit Commitment is a computationally demanding mixed-integer linear optimisation problem requiring many binary commitment decisions across a scheduling horizon. To reduce this computational burden, machine learning approaches can predict a subset of these decisions, thereby shrinking the search space explored by the optimisation solver. Existing confidence-based approaches determine which variables to fix based on user-defined probability thresholds, which are agnostic to the downstream optimisation problem. We address this limitation by introducing an optimisation-aware framework that yields generator-specific confidence thresholds based on the impact of fixing errors on Unit Commitment solution quality. This approach won first place in the 2025 EPRI AI-ccelerating Unit Commitment competition. The proposed framework achieves a mean optimality gap below 0.5% while also delivering an average speed-up of more than 20x. It therefore provides a general mechanism for translating probabilistic predictions into controlled reductions of mixed-integer search spaces, directly linking learning decisions to their downstream computational and economic consequences.
Read the full abstract ▾
All authors Benjamin Fritz, Andreas Makrides, Maryam Fetanat, Pierre Pinson
Information Theory · Cross-listed from quant-ph
Logical Operator Decomposition for Distance Analysis of Bivariate Bicycle Codes
Mohammad Rowshan and Simon Devitt
Bivariate bicycle (BB)
quantum codes are a prominent finite-length family of
quantum low-density parity-check codes, but their minimum distance is usually established numerically rather than read from the defining
polynomials. We study the $Z$-logical quotient $K/S$ over $\mathbb F_2[x,y]/(x^\ell-1,y^m-1)$ and show that it fits into a short exact sequence with an annihilator quotient as kernel and a colon quotient as cokernel. The sequence gives an explicit logical basis, a dimension formula, and a componentwise distance identity $d_Z=\min(d_{\mathcal A},d_{\mathcal C})$. Using the Frobenius structure of the finite group
algebra, we prove $r_{\mathcal A}=r_{\mathcal C}=k/2$ for every BB code, including repeated-root cases. The algebraic component of a logical class is distinct from the support shape of its lightest representatives: an annihilator class can have a lighter two-block representative, and a colon class can have a one-sided minimum. For lower bounds we show that every proper subset of a minimum-weight logical operator has nonzero syndrome, and that this property persists inside the colon component but not inside the annihilator component. A translation-anchored cluster search built on it proves the distances $4,6,10,10,12,18$ of the six standard BB codes of lengths $18$ to $288$ and enumerates every minimum-weight logical operator. The resulting census shows that $[[108,8,10]]$ is the only one of the six whose distance is attained in a single component, with $d_{\mathcal C}=10$ and $d_{\mathcal A}=12$.
Read the full abstract ▾
Artificial Intelligence
Can Computation from Earlier Problems Help LLMs Solve New Ones?
Jipei He, Wenhui Tan, Xiaoyi Yu et al. (+4 more)
Large
language models often solve independent problems in the same conversation. Can computation from earlier problems help them solve new ones? To answer this question, we first conduct preliminary experiments showing that retained history can raise or lower later-turn accuracy, even within the same domain. To understand these effects, we use controlled replay to isolate internal state changes specific to each problem-history pairing. Across different histories, these changes preserve similar relationships among current problems. To improve
reasoning under retained history, we introduce STAIR (Stale-
Token Attention for Inter-query Reuse). STAIR captures keys and values from earlier response generation in a fixed bank. It learns to redirect current queries when they read this bank during
prompt processing. The base model remains frozen; only 12,288 parameters are trained. Across three Qwen models and four
benchmarks, STAIR improves average later-turn accuracy by up to 11.67 percentage points over the unmodified model with history.
Read the full abstract ▾
All authors Jipei He, Wenhui Tan, Xiaoyi Yu, Enver Sangineto, Fiorenzo Parascandolo, Rita Cucchiara, Ruihua Song
Artificial Intelligence
Experimental Experience Modeling for Autonomous Research
Wenda Wei, Yingchen Zhang, Ruqing Zhang et al. (+3 more)
Autonomous research
agents can generate hypotheses and conduct experiments, but experimentation remains a major source of computational cost. A fundamental challenge is deciding which experiments are worth running, particularly when prior evidence is insufficient to resolve uncertainty. Yet current research
agents lack a systematic way to leverage experimental experience when making such decisions. We introduce Experimental Experience Modeling (EEM), a framework for making informed experimental decisions by acquiring, reusing, and accumulating experimental experience. EEM extracts decision-relevant records from earlier experimental trajectories, distills them into reusable experience, and organizes them in an experience library. For a new experimental decision, EEM retrieves relevant historical experience and assesses whether it provides sufficient support for deciding whether a candidate direction warrants further investment. When historical experience is insufficient, EEM conducts a targeted, low-cost pilot experiment to acquire the missing decision-relevant experience on demand. It then combines this newly acquired experience with retrieved historical experience to determine whether the direction warrants full-scale evaluation, which requires substantial resources. The resulting experimental outcomes are further distilled into reusable experience, allowing the library to continually grow through iterative accumulation. Experiments on autonomous research
benchmarks show that EEM improves research performance while reducing model interaction overhead, demonstrating the value of reusing accumulated experience and acquiring additional experience only when needed.
Read the full abstract ▾
All authors Wenda Wei, Yingchen Zhang, Ruqing Zhang, Jiafeng Guo, Daiting Shi, Xueqi Cheng
Machine Learning
Decoupled and Distilled: Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer for Few-Shot Class-Incremental Learning
Hongwei Zhao, Rui Liu, Yansong Liu et al. (+2 more)
Few-Shot Class-Incremental Learning (FSCIL) addresses the challenge of learning new classes from very limited samples while retaining knowledge of previously learned ones. Although parameter-efficient fine-tuning methods with pre-trained models show promise for class-incremental learning, strict gradient-based constraints can be unreliable under severe data scarcity, while multi-expert approaches can impose substantial inference-time costs. We propose TALON (Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer), an inference-efficient FSCIL framework. TALON dynamically allocates an independent LoRA-Teacher to each incremental task for task-specific representation learning, then distills multiple frozen teachers into a unified LoRA-Student through Ensemble Knowledge Transfer, eliminating runtime module selection or generation. A semantic-guided
distillation strategy weights teacher contributions by feature-space similarity to mitigate catastrophic forgetting and overfitting. Across three class-order runs, TALON achieves comparable or better mean average accuracy across four FSCIL
benchmarks, obtaining 86.68 +/- 1.22% on CUB200, 90.39 +/- 0.27% on CIFAR100, 78.38 +/- 0.94% on ImageNet-R, and 96.34 +/- 0.33% on miniImageNet. TALON uses up to 33x fewer deployment parameters and reduces average inference time per task to 26.7 s, a 41.70% reduction relative to ASP.
Read the full abstract ▾
All authors Hongwei Zhao, Rui Liu, Yansong Liu, Zhiyuan Zou, Yong Chen
Robotics
UniWAM Technical Report: Unified Mobile Manipulation via Mixed-Stream World-Action Modeling and Manipulation Anchor Pose Supervision
Wei Xue, Keliang Liu, Mingzhang Cui et al. (+14 more)
Mobile manipulation requires precise
navigation to a manipulation-ready pose followed by reliable object interaction. These two stages differ in action spaces and visual requirements, which complicates unified policy learning. In addition, collecting diverse real-world
navigation data with explicit manipulation-ready pose supervision remains costly and difficult to scale. We introduce UniWAM, a unified mixed-stream world-action model with separate action encoders and output heads for
navigation and manipulation, sharing a common backbone. This design supports joint representation learning on independently sampled
navigation and manipulation data. UniWAM supports independent inference for either stream and batch-parallel inference for both. We further introduce Manipulation Anchor Pose (MAP) supervision for where to stop and how to orient for manipulation. An automated pipeline constructs MAP-Data from large-scale 3D scenes, yielding over 1.5 million episodes and 7,500 hours. MAP-Data provides per-frame target-object bounding boxes and image-plane MAP coordinates as auxiliary
navigation supervision. Together with projected end-effector trajectories for manipulation, these prediction targets provide stream-specific image-plane supervision for action learning from
egocentric observations. With large-scale MAP-Data, UniWAM outperforms the strongest external baselines on our MAP-Bench by 30.1% in position error and 44.0% in heading error. Across 24 real-
robot tasks, UniWAM achieves leading results in MAP
navigation and mobile manipulation, with competitive manipulation performance. We have released code, data, and
benchmark.
Read the full abstract ▾
All authors Wei Xue, Keliang Liu, Mingzhang Cui, Jinhua Xie, Jinjie Wei, Jianan Hou, Jingcheng Lu, Lintao Wang, Kaixiang Qiu, Yizhou Liu, Xinghai Ye, Jinghang Han, Mingcheng Li, Jie Gu, Shunli Wang, Lihua Zhang, Dingkang Yang
Machine Learning
Beyond Simulation: Retain-and-Repair Neural Operators for Real-World Adaptation
Woojin Cho and Junghwan Park
Neural operators increasingly benefit from pretraining on numerical simulations, yet adapting them for real-world prediction remains challenging. We introduce the Retain-and-Repair Neural Operator (R$^2$NO), a framework for adapting simulation-pretrained operators to real-world data while retaining useful pretrained structure. The pretrained operator is first finetuned on real data and then frozen to provide a source prediction, and a shared repair module learns a sequence of refinements from the same observations. Using orthogonal
Fourier projections, a spectral ensemble fits a small ridge regression within each
cell of the
Fourier domain and combines the refinements by weights fitted on a held-out split of the real data. The
cells are defined jointly by radial ranges, angular sectors, and measured channels, allowing refinement depth to vary with frequency magnitude, with orientation, and across channels. Including the source prediction as a candidate makes retention available in every
cell, and independently trained repair modules enter the same combination as additional candidates. On all RealPDEBench systems and six backbones, R$^2$NO consistently outperforms full finetuning and iterative refinement. The framework treats adaptation depth as a
cell-specific choice learned from real data.
Read the full abstract ▾
Machine Learning
When, Not How Much: Evaluating Time-Series Foundation Models on Sparse Events
Daniel Schoess and Florian von Wangenheim
Pretrained time-series foundation models (TSFMs) are evaluated as forecasters of future values, yet for sparse series many decisions depend only on which future periods contain activity. Standard
benchmarks do not assess this. On five sparse
datasets, we rank positions within forecast windows that contain both events and zeros. The released point forecasts of 12 TSFMs improve chance-corrected average precision over training-free references by at most 0.031, and in chance-corrected AUC the median TSFM falls below them on every
dataset. With event supervision, linear probes of six frozen backbones improve on their backbone's point forecast in 29 of 30 backbone–
dataset pairs. Averaging the predicted
quantiles instead of taking their median improves the ranking of most TSFMs that forecast the median, and on two
datasets the strongest such outputs rival the probes. The probes' advantage over raw-context learners depends on the
dataset, and under the same probe, pretrained features outperform randomly initialized ones for five of six backbones. For sparse-event ranking, released point forecasts thus add little over simple references, whereas lightweight event heads on frozen TSFMs rank events better than these forecasts, and the best of them exceed gradient-boosted trees trained on the raw context on three of the five
datasets. More broadly, assessing pretrained forecasters on tasks beyond value
forecasting requires reporting their outputs, supervised probes of their representations, and raw-context and randomized controls side by side, since each supports a different conclusion.
Read the full abstract ▾
Computation and Language
TTLab at Daleel 2026: STAR-Ar, Sequence Tagging for Argument Recognition in Arabic
Bhuvanesh Verma, Ali Abusaleh and Alexander Mehler
Argument Mining (AM) is a critical NLP task that remains significantly under-resourced in Arabic. This paper presents $\testtt{STAR-Ar}$, a BERT-BiLSTM-CRF architecture for argument discourse detection and classification, as our system for Daleel 2026, the inaugural Arabic argument mining shared task. The task requires the identification and classification of argumentative discourse units (ADUs) in debate and editorial texts.We jointly model these two objectives as a
token-level sequence labeling task using a BERT-BiLSTM-CRF architecture that combines contextual
transformer embeddings with structural transition constraints to support accurate span detection. $\testtt{STAR-Ar}$ achieves an F1-score of 72.69 on validation and 73.7 on test data. Our domain-specific analysis shows that models trained exclusively on editorials underperform those trained on debates, a disparity we primarily attribute to the smaller size of the editorial
dataset. The code for $\testtt{STAR-Ar}$ is available at ${\href{https://github.com/ENTAILab/daleel_2026_Arabic-Argumentative-Discourse-Mining}{\faGithub~TTLab at Daleel 2026}}$
Read the full abstract ▾
Robotics
RoboAssist: Interactive Human-Humanoid Planning for Long-Horizon Surgical Assistance
Jingwei Jia, Keyu Zhou, Jiewei Wang et al. (+7 more)
Long-horizon
surgical assistance requires
humanoid robots to coordinate with evolving human activities while maintaining safety across planning and execution. We present RoboAssist, an
agent-based framework for interactive human-
humanoid planning that integrates workflow
reasoning, task coordination, and cross-layer safety. At its core is an asymmetric dual-track representation that separates partially observed human process states from executable
robot task sequences. By updating human-process estimates, scene context, and task dependencies online, RoboAssist revalidates the remaining task sequence and replans only the affected suffix when workflow requests change. A cross-layer safety architecture combines preventive
navigation regulation, reactive regulation during close-range handover, and independent whole-body runtime supervision. This design couples online task coordination with safety constraints throughout execution. We demonstrate the framework on a Unitree G1
humanoid robot in long-horizon, multi-stage simulated
surgical assistance scenarios encompassing
multimodal interaction, instrument handling, medical material transport,
navigation, and safe human-
robot handover. Experiments show multi-stage task completion and adaptation to workflow-request changes. A targeted full-replanning ablation shows that residual replanning reduces plan-update latency and post-update
token usage. Separate safety experiments demonstrate complementary protection across
navigation, handover, and runtime supervision. Additional results and demonstrations are available online at https://roboassist.github.io.
Read the full abstract ▾
All authors Jingwei Jia, Keyu Zhou, Jiewei Wang, Peisen Xu, Xingyuan Zhou, Liang Wang, Jiming Chen, Gaofeng Li, Jin Wang, Shunlei Li
Machine Learning
From Search to Signal: Online Post-Training in Automatic Heuristic Design
Yilun Yuan, Tianyu Zhou and Zhenzhou Tang
Large
language model (LLM)-based automatic heuristic design (AHD) iteratively proposes and refines heuristics, pairing design rationales with executable code. Task-specific evaluators assess programs; execution outcomes and performance scores guide search. Many AHD systems keep the generator frozen; EvoTune and Co-
Evolution of
Algorithms and
Language Model (CALM) instead update it from evaluated candidates. When such outcomes drive reinforcement learning with verifiable rewards (RLVR), they create a search-coupled loop: the evaluated candidate stream supplies both search-state updates and training signals for the model that generates future candidates. Yet validity and performance do not uniquely determine useful model updates; converting them into learning signals must account for the
prompt and evolving search state that produced each candidate. We formulate online post-training of small open-weight LLMs in AHD as context-dependent signal construction and develop alternative mappings from program validity, task performance, and generation context to update signals. Using shared evaluated rollouts and matched update budgets, controlled experiments across AHD tasks and model families compare these mappings with online post-training baselines, testing their effects on validity, performance among valid proposals, and the yield of valid proposals that improve under contextual comparisons. Complementary checkpoint, frozen-search, and live-system evaluations assess whether proposal-level gains appear in updated checkpoint behavior and subsequent search, rather than arising solely from accumulated search state. A resource-matched comparison under pre-specified cost accounting tests whether online updating adds value beyond additional search with a frozen generator. Together, this design avoids treating end-to-end search gains alone as evidence of stronger heuristic-design capabilities.
Read the full abstract ▾
Artificial Intelligence
SkillFM: Generating Skills for LLM Agents via Latent Flow Matching
Zuming Zhang, Jie He, Yizhe Zhang et al. (+1 more)
Textual skills provide reusable guidance for large
language model
agents, but existing approaches often rely on manually curated skill banks or reinforcement learning with indirect and delayed feedback. We introduce SkillFM (Skill Flow Matching), a generative framework that synthesizes task-conditioned textual skills directly without test-time skill
retrieval. Our framework combines a codec for encoding and reconstructing textual skills in a continuous latent space with a conditional flow model trained using improved MeanFlow. At inference time, the learned velocity field enables single-step latent sampling, and an LLM-based decoder converts the sampled representation into textual guidance for a frozen downstream
agent. We evaluate the framework on
embodied tasks, question answering, and web shopping. On ALFWorld and Search-QA, our method achieves the best overall performance among the compared vector-based skill approaches. Our analyses further demonstrate that latent skill generation is an effective alternative to
retrieval-based skill augmentation. Our code and training skill libraries are available at https://github.com/lulushang999/SkillFM.
Read the full abstract ▾
All authors Zuming Zhang, Jie He, Yizhe Zhang, Jeff Z. Pan
Computer Vision
InfoAgent: Traceable Generation and Repair of Evidence-Grounded Infographics
Yifan Li, Tong Li, Qi Zeng et al. (+6 more)
Reliable infographic generation requires facts, symbols, and visual relations to remain consistent through rendering and revision. Correcting one element also requires tracking its supporting evidence and the dependencies affected by the change. We present InfoAgent, a training-free framework for evidence-bound visual-symbolic program synthesis. Its Infographic Visual Description (IVD) records factual payloads, evidence provenance, execution routes, and verification obligations in a typed dependency graph. Retrieved design priors guide compilation, and layered execution combines raster synthesis with editable symbolic and binding objects while retaining their traces. Dependency-aware repair localizes corrections, rechecks affected dependencies, and requires protected obligations to remain satisfied under the declared checkers. Unresolved obligations remain explicit. On IGenBench, InfoAgent achieves 93.0 Q-ACC and 59.0 I-ACC. We also introduce InfoGraphicBench-Evidence, where complete-checklist pass rates on 200 test requests increase from 21.5% for Same-IVD
Prompt to 23.5% for the initial layered output and 28.5% after repair, using the same evidence and initial IVD. On 120 audited repair cases, localized repair edits 12.4% of the canvas on average, compared with 67.3% for global regeneration.
Read the full abstract ▾
All authors Yifan Li, Tong Li, Qi Zeng, Lishuai Gao, Ruwei Pan, Cong Wei, Shaohua Kevin Zhou, Zhuoliang Kang, Xiaoming Wei
Computer Vision
Shulin Tian, Junsu Kim, Shuai Liu et al. (+17 more)
Real-world
embodied tasks, from everyday activities to professional procedures, require
agents to act under physical constraints while tracking evolving object and task states. Tool use sits at the heart of such tasks, as many everyday and professional activities are tool-mediated. Understanding them requires
reasoning about affordances, hand-tool-object geometry, procedural progress, and causal effects on target objects. Yet despite strong performance on perception-oriented
video tasks such as captioning and general
video QA, current
multimodal video models remain limited in this form of tool-centric
embodied reasoning. Progress in this direction has been limited by the lack of real-world
egocentric data and diagnostic
benchmarks. To address this gap, we introduce EgoTools, the first comprehensive suite for
egocentric tool-use understanding. It consists of two complementary components: EgoTools-Data, a large-scale corpus of 100 hours of tool-centric
egocentric recordings with synchronized audio, dense captions,
reasoning-heavy narrations, and supplementary 3D information; and EgoTools-Bench, a diagnostic
benchmark of 1,000 QA pairs across four tracks that cover tool-use understanding from perception and geometry to procedure and causal
reasoning. Experimental results show that current models still struggle to ground tool use in visual evidence: Gemini-3.1-Pro achieves 66.9% overall accuracy but only 51.7% on Perception & Grounding. Beyond evaluation, we validate EgoTools-Data as a training resource. On the full 1,000-question
benchmark, full supervised fine-tuning improves Qwen3-VL-8B-Instruct from 50.0% to 60.9%, under strict source-
video separation. Together, these results establish EgoTools as a unified resource for both training and diagnostic evaluation of real-world
egocentric tool-use understanding.
Read the full abstract ▾
All authors Shulin Tian, Junsu Kim, Shuai Liu, Hao Li, Yujiao Shen, Sihan Li, Zhe Yang, Yeongon Kim, Feiyu Li, Jialin Wu, Yichi Zhang, Wenhui Wang, Runmao Yao, Yuhao Dong, Zhaoxi Chen, Fangzhou Hong, Antonino Furnari, Jingkang Yang, Hongyuan Zhu, Ziwei Liu
Robotics
Beyond the Current Scene: Event-Referential Grasping with Active View Selection
Hyunjoon Lee, Haebeom Jung, Eunsung Cha et al. (+4 more)
A
robot that observes people interacting with objects should be able to carry out later requests that refer back to those interactions. Such requests may specify a grasp target by the role it played in a past event rather than by its name or appearance. Moreover, the target may no longer be visible when the
robot is asked to act. We present BeyondSCe, a zero-shot robotic grasping system for this event-referential setting. Given the event history and the current scene, the system identifies the requested object or part and localizes it for grasping. If the target is occluded, it combines an event prior recovered from the history with current scene geometry to select camera viewpoints likely to reveal the target. The system uses pretrained models without additional task-specific training. In real-
robot experiments with a single wrist-mounted RGB-D camera, it achieves grasp success rates of 76% and 77% for initially visible and occluded targets, respectively, compared with 40% and 55% for the strongest baseline in each condition. On four additional scenes with heavy occlusion, it increases grasp success rates from 75% to 95% while reducing the mean number of views from 3.35 to 2.20, compared with an active-perception baseline given the target's ground-truth 3D bounding box.
Read the full abstract ▾
All authors Hyunjoon Lee, Haebeom Jung, Eunsung Cha, Daeun Lee, Yu-Chiang Frank Wang, Jaesung Choe, Jaesik Park
Machine Learning
Wavelet Flow Matching for Time Series
Lucas Poinsignon, Jorge da Silva Goncalves, Samuel Ruiperez-Campillo et al. (+1 more)
Synthetic time series are increasingly used for data augmentation,
privacy-preserving data sharing, and downstream model development, yet faithfully reproducing both multi-scale temporal structure and cross-channel dependencies remains challenging. We study multivariate time-series generation through flow matching in the wavelet domain. By operating on multilevel discrete wavelet coefficients rather than directly in the time domain, the model represents coarse structure and progressively finer details at separate scales. Their naturally different variances further induce an implicit coarse-to-fine generative process without requiring an explicit multi-scale schedule. Since the transform acts independently on each channel, we pair it with a channel-
token transformer whose
attention directly models cross-channel dependencies. Across seven
benchmark datasets and four sequence lengths, our method is best or tied on a majority of
dataset-metric combinations, with the largest and most consistent improvements in Context-FID and discriminative score.
Read the full abstract ▾
All authors Lucas Poinsignon, Jorge da Silva Goncalves, Samuel Ruiperez-Campillo, Julia E. Vogt
Information Theory
Coded Computing for Dynamic System via Cartesian Products
Chenglin Li, Minquan Cheng and Youlong Wu
Coded distributed computing (CDC) mitigates the Shuffle communication bottleneck through coded multicast, but existing schemes with predetermined worker populations do not directly adapt to worker arrivals and disconnections. To address these dynamics, we propose a dynamic CDC scheme that preserves surviving workers' caches and Reduce assignments while exploiting the storage brought by arriving clusters. The model accommodates general MapReduce tasks under a placement that is partly fixed and partly new. The scheme uses Cartesian products of cluster-wise placement delivery arrays (PDAs) for file placement. A delivery-aware greedy
algorithm reassigns abandoned Reduce functions, and a Communication Array allows workers with no Reduce function to serve as coded transmitters. For arbitrary feasible disconnections, the scheme's exact communication load is derived. A file-wise converse under one-shot communication and uncoded placement shows that, in the absence of disconnections, the scheme lies within a factor of two of the lower bound under these assumptions when new clusters arrive, and within a factor of four otherwise.
Read the full abstract ▾
Artificial Intelligence
EHR-RobustGym: Benchmarking and Training Agents for Robust Clinical Reasoning
Yitong Qiao, Yancheng Jin, Lei Liu et al. (+4 more)
In hospital workflows, electronic health records (EHRs) are often noisy, and may not contain the evidence needed to confirm events or measurements referenced in a clinical query. Even when database
retrieval succeeds, clinical
agents can overlook such discrepancies and return plausible but unsupported answers. We introduce EHR-RobustGym, a scalable and interactive environment for evaluating and training robust clinical
agents grounded in noisy EHRs. Built on MIMIC-IV hospital records (365K patients, 31 tables, and over 500M records), EHR-RobustGym comprises 5,486 Clean-Noise pairs spanning six clinical intents and both patient-level and population-level queries. The pairs test robustness to Record-level, Value-level, and Query-level noise, while interactive SQL/Python execution and outcome verification support trajectory collection and training. Evaluating multiple LLMs reveals substantial robustness gaps: average task success across proprietary and large-scale open-weight models drops from 62.2% on Clean questions to 37.9% on Noise questions. At k=4, pass^k consistency falls below 50% for most evaluated models, exposing instability in clinical task completion. Supervised fine-tuning and reinforcement learning in EHR-RobustGym improve performance, with gains generalizing to five external EHR
benchmarks. Together, these results position EHR-RobustGym as a testbed for evaluating and improving the evidence-grounded robustness of clinical
agents.
Read the full abstract ▾
All authors Yitong Qiao, Yancheng Jin, Lei Liu, Yue Shen, Jian Wang, Jinjie Gu, Zhixuan Chu