Numerical Analysis
Constraint Preserving AFD-WENO Schemes for Relativistic Hydrodynamics with General Equations of State
Pramodit Mishra, Shubham Upadhyay, Rakesh Kumar et al. (+1 more)
We develop a high-order physical-constraint-preserving (PCP) alternative finite difference weighted essentially non-oscillatory (AFD-WENO) scheme for the special relativistic hydrodynamics equations with general equations of state. The proposed scheme comprises two key limiters: a state limiter, which acts after the WENO state interpolation step, and a flux limiter, which acts on the final high-order fluxes. The state limiter ensures that the interpolated states are physically admissible, while the flux limiter ensures that the numerical fluxes are physically admissible. The resulting scheme is rigorously proved to satisfy the physical constraints. Incorporating multiple WENO interpolation techniques, including an improved adaptive-order formulation (WENO-AOI), the method is validated through extensive one- and two-dimensional numerical
benchmarks with various equations of state. The numerical results demonstrate high-order accuracy, sharp resolution of discontinuities, and robust stability in extreme relativistic regimes.
Read the full abstract ▾
All authors Pramodit Mishra, Shubham Upadhyay, Rakesh Kumar, Biswarup Biswas
Information Retrieval
ICEGR: An Intent-Coherent End-to-End Generative Retrieval Framework for E-commerce Search
Jiayi Tuo, Hehan Li, Dongjun Fu et al. (+10 more)
Generative
Retrieval (GR) is promising for e-commerce search, yet existing methods struggle to maintain query-intent consistency throughout the training pipeline. First, semantic ID (SID) construction based on static product information limits the ability of SIDs to encode product-intent associations. Second, although supervised fine-tuning (SFT) learns product-SID mappings across the catalog, low-exposure products still lack real query-intent supervision because query-to-SID training relies solely on online logs, resulting in poor
retrieval performance for these products. Third, business-oriented preference optimization may favor popular or high-value products over those that best match the query intent, weakening query-product relevance. To address these issues, we propose ICEGR, an Intent-Coherent End-to-End Generative
Retrieval Framework for E-commerce Search that integrates query intent consistently throughout the GR training pipeline. ICEGR comprises three components: (1) Intent-Aware SID Construction incorporates query-intent signals into SID construction, enabling SIDs to capture search intent beyond static product information; (2) Synthetic Query-Enhanced Unified SFT unifies multiple SFT tasks under the query-to-SID objective and augments sparse supervision from online logs with synthetic queries, providing complementary query-intent supervision for low-exposure products; and (3) Relevance-Calibrated Preference Optimization integrates query-product relevance and business signals into a margin-adaptive preference objective, preserving query intent while enabling business preference learning. Offline results show that ICEGR improves Recall@20 by 21.7% and NDCG@20 by 26.6% over the baseline. Deployed as an end-to-end generative
retrieval pathway in Baidu E-commerce Search, ICEGR achieves relative improvements of 3.52% in CTR, 15.96% in order volume, and 7.53% in GMV in an A/B test.
Read the full abstract ▾
All authors Jiayi Tuo, Hehan Li, Dongjun Fu, Xin Lu, Ling Zhuang, Fuwei Zhang, Meifang Li, Peizhi Xu, Hanmeng Liu, Shuanglong Li, Liwei Qian, Yanbiao Ma, Fuzhen Zhuang
Systems and Control
On the Number of Observation Nodes in Recurrent Neural Networks with Linear Threshold and ReLU Functions
Liangjie Sun, Wai-Ki Ching and Tatsuya Akutsu
This paper investigates how node update rules and admissible state domains affect the minimum number of observation nodes required for global finite-horizon observability in recurrent neural networks with linear-threshold and ReLU update functions. Over a common binary state domain, we construct a class of $K$-linear-threshold ($K$-LT) networks whose initial states can be uniquely reconstructed from the finite output trajectory of a single observation node. We further establish a dynamical equivalence between binary-valued $K$-ReLU networks and $K$-AND Boolean networks, which transfers existing class-level observation-node bounds to binary-valued $K$-ReLU networks. For nonnegative-valued ReLU networks, the observability problem reduces to the classical linear-system setting whenever the relevant pre-activations remain nonnegative. For general real-valued ReLU networks, we prove that global finite-horizon observability requires at least $n/2$ observation nodes when no restriction is imposed on the number of state variables involved in each node update. This lower bound is tight when $n=2K$, for which we construct a $K$-ReLU network observable from exactly $K$ nodes. These results show that both update rules and state domains fundamentally affect extremal observation requirements: temporal
evolution can concentrate finite-state information into a single measured trajectory, whereas activation-induced rank loss creates an intrinsic sensor lower bound in continuous-state ReLU networks.
Read the full abstract ▾
Machine Learning
Reward-guided Fine-Tuning of One-Step Generative Models via Wasserstein Gradient Flow
Hoseong Hwang, Woorim Han, Joungin Chun et al. (+2 more)
To mitigate the time complexity of generative models, one-step generative models have recently emerged through direct mapping from noise to data in a single forward pass. However, the reward-guided fine-tuning method of one-step generative models remains largely unexplored. To address this, we consider one-step generators from an optimal transport view, investigating Wasserstein Gradient Flow (WGF) for modeling smooth and controlled distributional
evolution in probability space. We then propose a novel reward-guided fine-tuning of a one-step generative model via WGF. We derive a practical training method that requires no reward gradients, thereby handling both non-differentiable and differentiable rewards. Moreover, our method provides smooth and stable reward-guided distributional updates while mitigating reward hacking and mode collapse. Experiments on 2D synthetic data, CIFAR-10, and ImageNet 256$\times$256 with diverse rewards, including JPEG (in)compressibility, class probability, Black-and-White and CLIP
alignment, show that our method achieves better reward
alignment compared to baselines.
Read the full abstract ▾
All authors Hoseong Hwang, Woorim Han, Joungin Chun, Jinseong Park, Jaewoong Choi
Artificial Intelligence
Cascade Failure Attribution for Multi-Agent Systems
Jiayi Zhang, Zexin Wang, Degang Sun et al. (+4 more)
Large
language model (LLM)-based
agents have shown strong potential in solving complex tasks through multi-step
reasoning, yet they remain vulnerable to execution failures. Accurate failure attribution is therefore critical for improving
agent reliability. Existing
topology- and spectrum-based methods exploit trajectory structures but often overlook fine-grained semantics, while LLM-based attribution methods capture semantic cues but suffer from long-context degradation over lengthy trajectories. To address these challenges, we propose DuoTrace, a plug-and-play detection filter for LLM-based failure attribution. DuoTrace follows a detect-before-attribute paradigm: it first detects anomalous executions and then supplies focused trajectory evidence to downstream LLM-based attribution methods. For effective VAE-based anomaly detection on
agent trajectories, DuoTrace integrates dual-view semantic-structural node representations, a Tree-LSTM-based trajectory encoder, and data augmentation through outcome-verified LLM replay and root-to-node path extraction to handle heterogeneous nodes, hierarchical execution structures, and limited normal execution data. Experiments with six LLM-based attribution baselines show that DuoTrace improves
agent-level and step-level attribution accuracy by 11.2% and 7.7%,respectively.
Read the full abstract ▾
All authors Jiayi Zhang, Zexin Wang, Degang Sun, Changhua Pei, Fei Sun, Gaogang Xie, Jingjing Li
Computer Vision
Conducting Stylistic Analysis of Paintings through an Art-History Agent
Marc S. Walton and Astrid Harth
Attributing an artwork to an artist has traditionally relied on detailed visual observations and descriptions, known as stylistic analysis in art history. By contrast, current artificial intelligence (AI) models used in the field offer only unexplained probabilistic classifications. To bridge this methodological gap, we present an AI framework that automates stylistic analysis of paintings, providing a foundation for enhancing evidence collection, discovery, and verification. By training a
vision transformer (ViT) on a large corpus of paintings with metadata, our system encodes this art history-specific data as
embeddings. These representations are factorized via sparse dictionary learning into a shared set of features that recur across the training set. A large
language model (LLM) then interprets each feature by retrieving associated artworks and their accompanying curator-written texts, and synthesizes them into descriptions that reflect their stylistic attributes. Finally, an autonomous coordinator LLM applies a
reasoning-and-action (ReAct) framework to weight, test, and refine these features into cohesive descriptions of an artwork, or comparisons of artworks. This approach converts detailed visual features into descriptive terms, addressing a key challenge in art history. It thus connects the use of images as data with the semantic concerns of humanists, establishing
vision-based computational art history as an area for future growth.
Read the full abstract ▾
Multiagent Systems
Harness-RL: Black-Box Reinforcement Learning with Action-Args Decoupling for Central-Agent Multi-Agent Harnesses
Xinke Jiang, Zhixin Zhang, Zhibang Yang et al. (+6 more)
Large
language model
agents increasingly solve long-horizon tasks through multi-
agent harnesses in which a central
agent coordinates specialized sub-
agents, tools, and environments. Training the central policy in such a harness raises two challenges. First, an action label is a low-cardinality decision, whereas its args form a high-dimensional conditional sequence; optimizing both with a shared sequence-level signal can produce conflicting gradients. Second, dynamic scheduling creates interdependent sessions with branches, parallel calls, and rewritten contexts, which cannot be faithfully reduced to one flat
token sequence. We introduce Harness-RL, a structured reinforcement learning framework that combines Conflict-Aware Policy Optimization (CAPO) with interface-level black-box trajectory construction. The black-box component captures Interface Call Records, builds per-session prefix trees, and aligns outcome and process rewards with trainable
tokens. CAPO uses forward activations to identify parameter partitions associated with action and args
tokens, then routes their policy gradients to the corresponding subspaces. Harness-RL supports both central-only and joint multi-
agent training. Across seven multi-hop question answering and agentic
retrieval benchmarks, it reaches average F1 scores of 42.93 and 47.79 with Qwen2.5-1.5B and Qwen2.5-3B, respectively, while ablations validate the contribution of CAPO and favor central-only optimization in the evaluated setting. Our code is available at https://github.com/jiangxinke/Harness-RL.
Read the full abstract ▾
All authors Xinke Jiang, Zhixin Zhang, Zhibang Yang, Jiaran Gao, Rihong Qiu, Shijin Chen, Xu Chu, Junfeng Zhao, Yasha Wang
Machine Learning
LLMODE: Aligning ODEs with LLMs via Gated Token Injection for Irregular Spatio-Temporal Forecasting
Di Zhang, Jingyang Zhang, Ziqian Wang et al. (+4 more)
Large
language models (LLMs) have shown promise for spatio-temporal
forecasting, but existing approaches often rely on regularly sampled
token sequences and struggle with irregular observations because of temporal asynchrony, representation-space misalignment, and limited context windows. We propose LLMODE, a
token-efficient framework for irregular spatio-temporal
forecasting with a frozen LLM backbone. LLMODE first uses a graph-aware ODE encoder to reconstruct irregular graph observations as a continuous-time latent trajectory. A Fixed-Budget Perceiver Resampler then compresses this variable-length trajectory into a fixed number of dynamic
memory tokens. In parallel, compact statistical descriptors are encoded and resampled into context
memory tokens. A dual-source gated cross-
attention module injects both
memories into the frozen LLM, enabling controlled utilization of external spatio-temporal evidence. Experiments on three real-world urban
datasets and two physical-dynamics
benchmarks show competitive overall performance, with clearer advantages under sparse or dynamically complex irregular sampling. Additional evaluations on unseen urban regions further demonstrate strong zero-shot generalization without adaptation.
Read the full abstract ▾
All authors Di Zhang, Jingyang Zhang, Ziqian Wang, Chi Zhang, Yikun Ban, Ziwei Zhang, Ruijie Wang
Formal Languages and Automata
Finite-Monoid Compression in Syntactic Concept Lattices: Arity Hierarchies and a Pseudovariety Trichotomy
Takayuki Kuriyama
Clark's syntactic concept
lattice (SCL) records two-sided distributional structure, and Wurm extended it to tuples of arbitrary finite arity. We study \(\operatorname{cmp}_f(L)\), the minimum image size of a finite-monoid observation that preserves guarded tuple substitution through arity \(f\) on the principal layer. For regular
languages, we characterize \(\operatorname{cmp}_f(L)\) exactly as the least cardinality of the codomain of an \(f\)-separating relational morphism from the pointed syntactic monoid. Let \(\operatorname{ch}(\mathbf V)\) denote the least arity at which these compression numbers stabilize uniformly over a pseudovariety \(\mathbf V\). Our main result is the following trichotomy of possible uniform heights: \(\operatorname{ch}(\mathbf V)\in\{1,2,\infty\}\), with \(\operatorname{ch}(\mathbf V)=\infty\) if and only if \(\operatorname{Synt}(\{ab\})\in\mathbf V\). Thus no finite uniform compression height \(3,4,\ldots\) occurs. The infinite case is sharp: inside \(\langle\operatorname{Synt}(\{ab\})\rangle\), every boundary \(d\to d+1\) admits unbounded compression gaps, and arbitrary finite strict prefixes of the arity hierarchy are realizable. On the finite side, commutative monoids and bands stabilize at arity one, while every completely regular syntactic monoid stabilizes by arity two; finite group kernels show that the binary bound is sharp. At unary arity, every nonempty finite simple graph is realized by an explicit length-three
language, yielding an exact chromatic-number formula and NP-completeness of deciding \(\operatorname{cmp}_1(L)\le 3\) for explicitly listed length-three
languages. The structural boundary between compression heights one and two remains open.
Read the full abstract ▾
Numerical Analysis
Tensor Orthogonal Subspace Split: Theory and Applications
Jifei Miao, Juan Han, Michael K. Ng et al. (+1 more)
Tensor representations have emerged as a fundamental paradigm for modeling multidimensional data by preserving intrinsic correlations across multiple modes. This paper proposes a novel theoretical framework, termed Tensor Orthogonal Subspace Split (TOSS), which explicitly splits a tensor, along a prescribed mode, into two orthogonal components: a dominant component lying in a prescribed subspace and a residual component lying in the corresponding orthogonal complement. We first present the general formulation of TOSS and systematically investigate its fundamental properties. As an important and practically meaningful special case, we further introduce the rank-one TOSS, which imposes a separable rank-one structure along the splitting mode and admits a clear
geometric interpretation. This formulation naturally captures dominant consistent patterns while effectively isolating orthogonal residual component. The proposed framework establishes a unified theoretical foundation for tensor-domain orthogonal split and opens new avenues for structured tensor modeling across diverse applications. Building upon the developed TOSS theory, hyperspectral image restoration and color
video background modeling are considered as two representative tasks, for which corresponding optimization models are formulated. Efficient
algorithms are developed to solve the resulting problems. Extensive experimental results validate the effectiveness and superiority of the proposed approaches.
Read the full abstract ▾
All authors Jifei Miao, Juan Han, Michael K. Ng, Kit Ian Kou
Machine Learning
Unsupervised Multi-Scale Gromov-Wasserstein Hypergraph Alignment
Lutz Oettershagen, Honglian Wang and Aristides Gionis
We study unsupervised hypergraph
alignment, where the goal is to infer node correspondences between two hypergraphs using only structural information, without node features, labels, seed matches, or side information. Direct higher-order formulations can represent hyperedge interactions faithfully, but they can be computationally demanding and cumbersome for non-uniform hypergraphs. Graph-reduction approaches introduce a different challenge: clique expansions keep the
alignment problem on the original node set but collapse all hyperedge evidence into one pairwise graph, whereas bipartite expansions preserve incidence structure but enlarge the problem from nodes to nodes plus hyperedges. We introduce FALCON (Filtration-based hypergrAph
aLignment via Cross-scale Optimal traNsport), an unsupervised optimal-transport framework for hypergraph
alignment. Instead of representing each hypergraph by a single collapsed clique graph, FALCON constructs a filtration-induced sequence of clique-based co-occurrence dissimilarity
matrices and jointly aligns all levels through one shared multi-scale Gromov–Wasserstein (GW) objective. The shared transport plan enforces a globally consistent node correspondence across filtration levels while avoiding the auxiliary hyperedge nodes introduced by bipartite expansion. Experiments on perturbation
benchmarks derived from real-world hypergraphs show that FALCON is robust to structural noise and in almost all cases outperforms strong graph- and hypergraph-
alignment baselines.
Read the full abstract ▾
Computational Engineering
QoI-Aware Provisional Rollout and Retrospective Reconciliation for Reduced-State Scientific Twins
Liangji Zhu, Scott Klasky, Jaemoon Lee et al. (+3 more)
Scientific twins may need to continue operating when updates from an authoritative primary system are temporarily unavailable. Once synchronization resumes, the new boundary can also be used to revise the intervening history. We distinguish an immediately available causal provisional trajectory from a delayed, future-conditioned reconciled trajectory. For reduced-state twins, we introduce a deterministic, calibration-based reconciliation method. A smooth temporal bridge carries the residual observed at the next synchronization block backward through the provisional interval. An analytic energy-matching stage then applies smooth regional gains and a global rescaling to match a component-energy trajectory estimated by cubic regression in log-energy space from synchronized frames on both sides of the gap. The method uses no additional correction network and revises decoded history without changing the latent state used for later rollouts. We evaluate 64 spatial patches from 16 JHTDB isotropic-
turbulence slices for both velocity components and gaps S in {4, 6, 8}. During the longest gap, field error and gradient-sensitive QoI error degrade at markedly different rates, so field error alone does not characterize provisional fidelity. At S = 8, full reconciliation reduces window-averaged NRMSE by about 60% for both components and global gradient-intensity error from 4.21% to 2.91% for vx, whereas future-aware physical interpolation reaches 20.40% on the same metric. Energy matching additionally makes the reconciled history match its boundary-inferred global energy trajectory exactly. Future boundary information therefore substantially improves scientifically relevant properties within the evaluated regime.
Read the full abstract ▾
All authors Liangji Zhu, Scott Klasky, Jaemoon Lee, Qian Gong, Anand Rangarajan, Sanjay Ranka
Software Engineering
InteractBench: Benchmarking LLMs on Competitive Programming under Unrevealed Information
Jiaze Li, Aocheng Shen, Bing Liu et al. (+4 more)
Competitive programming is increasingly being used to evaluate the algorithmic
reasoning capabilities of large
language models (LLMs). However, existing
benchmarks primarily focus on full-information tasks where all problem inputs are provided upfront. This overlooks a critical dimension of algorithmic
reasoning: the ability of generated programs to operate when key information is not revealed upfront. Interactive problems, a distinctive component of competitive programming, embody this challenge. These problems require programs to engage in multi-round interaction with an interactor (a judge program) under strict protocol constraints and limited query budgets, with new information revealed only in response to queries. To address this gap, we introduce InteractBench, a
benchmark comprising 322 high-quality interactive problems curated from Codeforces, AtCoder, IOI, and ICPC. Each problem is packaged with executable local interactors, enabling fully offline evaluation. Unlike existing
benchmarks, InteractBench assesses whether model-generated code can acquire information and track state dynamically. Our evaluation reveals a significant interaction gap: even the most advanced
reasoning models achieve limited success on interactive problems. Beyond success rates, we propose a fine-grained failure taxonomy to diagnose the root causes of these deficiencies. Although algorithmic logic errors remain dominant, protocol violations and query-budget overruns are frequent. Code is available at https://github.com/kmsgk0/InteractBench.
Read the full abstract ▾
All authors Jiaze Li, Aocheng Shen, Bing Liu, Boyu Zhang, Xiaoxuan Fan, Qiankun Zhang, Xianjun Deng
Information Theory
Efficient Polynomial-Time Decoding of Simplicial Anticodes with Near-Optimal Performance
Antonio Jesús Lorite-López, Daniel Camazón-Portela and Juan Antonio López-Ramos
In this work, we propose an efficient decoding
algorithm for codes arising from simplicial complexes, a family of binary linear codes for which no decoding method of this type was previously known. Although the
algorithm does not always attain the maximum theoretical error-correcting capability, it provides an explicit bound that can be computed directly from the structure of the complex. Moreover, this bound is asymptotically optimal: the ratio between the guaranteed correcting capability and the theoretical maximum converges to $1$ as the code length increases, under natural assumptions on the dimension of the maximal faces. The correction capability is also presented in specific examples. Finally, we introduce specific families of simplicial complexes where the
algorithm successfully reaches this theoretical bound.
Read the full abstract ▾
Systems and Control · Cross-listed from math.OC
Reciprocal-Manifold Annealed KKT Flows for Constrained Optimization: Application to the Nonconvex AC Optimal Power Flow
M Parimi, Aditi Ramteke, Rachit Mehra et al. (+2 more)
Safety-critical optimization applications, such as real-time power system operation, maintain feasibility at every intermediate step, not merely at convergence. Existing approaches either violate constraints mid-solve (interior-point methods) or enforce feasibility through per-instant quadratic programming subproblems with cubic computational cost and unbounded worst-case execution time. We propose a continuous-time optimization framework for smooth constrained nonlinear problems that preserves feasibility throughout the optimization process without requiring projection operators, quadratic programming subproblems, or other per-iteration optimization routines. The method is built around a reciprocal multiplier
manifold, which establishes an explicit relationship between inequality constraints and their associated Lagrange multipliers. By designing a continuous multiplier update law, the
manifold is shown to remain forward invariant, while the resulting dynamics are equivalent to continuous-time logarithmic barrier gradient descent. The proposed framework naturally extends to multiple inequality constraints, equality constraints, nonconvex feasible sets, and infeasible initial conditions. The method is further enhanced through an augmented Uzawa flow that eliminates oscillatory transients commonly observed in classical primal-dual saddle-point dynamics. The effectiveness of the proposed approach is applied to the AC Optimal Power Flow problem of IEEE 9-bus and IEEE 57-bus systems. Numerical results show convergence to solutions within 0.4% of the
benchmark optimum while maintaining strict feasibility of all constraints. A computational complexity analysis shows that the proposed dynamics reduce the per-step computational cost from cubic to linear complexity. Finally, dynamic tracking studies under time-varying operating conditions demonstrate reliable feasibility preservation.
Read the full abstract ▾
All authors M Parimi, Aditi Ramteke, Rachit Mehra, Arun Mahindrakar, Navdeep Singh
Computer Vision
SPLG-Mamba: Structure-Preserving Local-Global Mamba Network for Salient Object Detection in Optical Remote Sensing Images
Yi Xu, Ruichao Hou, Tongwei Ren et al. (+1 more)
Salient object detection in optical remote sensing images (ORSI-SOD) requires dense predictions that preserve object completeness and structural continuity under complex backgrounds, scale variation, and irregular object shapes. Existing methods often localize salient regions, but their predictions may still suffer from structural degradation, including fragmented, incomplete, or locally missing foreground responses. This degradation is closely related to hierarchical feature propagation, where shallow details can introduce texture-induced background responses, deep semantics may over-smooth weak structures, and uncontrolled cross-scale fusion can disturb coherent regions. To address this issue, we propose a novel Structure-Preserving Local-Global Mamba Network, SPLG-Mamba, for ORSI-SOD. Specifically, SPLG-Mamba integrates Smooth-Detail Recalibration (SDR), hierarchy-aware Local-Global Mamba, and Gated Cross-Scale Fusion (GCSF). SDR recalibrates smoothed responses and detail residuals before state-space modeling, Local-Global Mamba assigns local modeling to shallow feature levels and global modeling to deep feature levels, and GCSF controls cross-scale detail injection during decoding. Experiments on ORSSD, EORSSD, and ORSI-4199 demonstrate state-of-the-art results and improved structural completeness and continuity. The code is available at https://github.com/yxu9910/SPLG-Mamba
Read the full abstract ▾
All authors Yi Xu, Ruichao Hou, Tongwei Ren, Gangshan Wu
Computation and Language
PrivBench: A Holistic and Modular Benchmarking Platform for Evaluating Text-to-Text Privatization
Stephen Meisenbacher, Andreea-Elena Bodea, Ahmet Bilal Ak{\i}n et al. (+3 more)
Natural
Language Processing methods have enabled novel solutions and advances in the field of
privacy, particularly in the sub-domain of text-to-text privatization, where the goal is to transform a sensitive input text into a privatized output by ideally masking (in)directly identifiable or otherwise private information. The evaluation of text-to-text privatization, however, is not straightforward, and the extant literature has utilized a myriad of techniques and metrics to quantify the
privacy-preserving capabilities of privatization methods. Seeking to unify the evaluation of text-to-text privatization, we introduce PrivBench, a holistic and modular benchmarking platform for researchers and practitioners working on text privatization. PrivBench is holistic in that it evaluates privatization on a series of defined desiderata, which are structured into modules. PrivBench is not only modular but also extensible, allowing for future updates and
benchmark versions. PrivBench is user-centered and promotes competition via real-time evaluation and a live public leaderboard. The platform is free to use and openly accessible at https://privbench.com/.
Read the full abstract ▾
All authors Stephen Meisenbacher, Andreea-Elena Bodea, Ahmet Bilal Ak{\i}n, Alexandra Klymenko, Jana Diesner, Florian Matthes
Computation and Language
MI-Distillation: Selecting from Model-Interpolated Instruct-Reasoning Data Spectrum for Chain-of-Thought Distillation
Yangsong Lan, Renkai Hu, HongKai Zheng et al. (+4 more)
Recent advances in large
reasoning models (LRMs) have shown strong performance on complex problems through long chain-of-thought (Long CoT)
reasoning. However, distilling such trajectories into smaller student models remains challenging: direct Long CoT supervision often provides limited gains and can be less effective than concise Short CoT rationales. In this work, we investigate this phenomenon from a gradient-centric perspective. Our analysis shows that Long CoT induces larger gradient magnitudes and more concentrated update directions than Short CoT, with this effect becoming more pronounced as student model capacity increases. These findings suggest that effective Long CoT
distillation requires balancing the
reasoning information density of
reasoning trajectories with their distributional
alignment to the student model. Motivated by this insight, we propose Model Interporlation
Distillation (MI-
Distillation), a framework that constructs a continuous Instruct-
Reasoning data spectrum through model interpolation. To select suitable trajectories from this spectrum, we further introduce Sequential Learnable Surprisal Score (SeqLSS), which favors
reasoning paths that are both informative and learnable for the student. Extensive experiments on
reasoning benchmarks show that MI-
Distillation consistently improves small model CoT
distillation over strong Long CoT baselines.
Read the full abstract ▾
All authors Yangsong Lan, Renkai Hu, HongKai Zheng, Bo Zhang, Renzhi Wang, Hongliang Dai, Piji Li
Multiagent Systems
AgenticRag-R1: Agentic Reinforcement Learning with Stack Memory for Multi-Step Reasoning, Retrieval and Memorizing
Xinke Jiang, Yue Fang, Zhibang Yang et al. (+12 more)
Retrieval-Augmented Generation (RAG) improves the factuality of large
language models (LLMs), yet existing RAG systems often struggle with complex, multi-step
reasoning that requires adaptive
retrieval and continuous revision of intermediate contexts. Recent reinforcement learning (RL)-based agentic RAG methods partially alleviate this issue, but typically rely on coarse-grained action spaces and trajectory-level rewards, resulting in weak reward assignment and a bias toward short-horizon, stereotyped
reasoning template. To address, we propose AgenticRag-R1, a RL framework that deeply integrates
reasoning,
retrieval, and
memory via a
memory stack and fine-grained action space, supported by hierarchical action-aware rewards and an information-aware trajectory rejection strategy to enable effective long-horizon learning. Experiments across a diverse set of multi-hop, open-domain, and agentic
reasoning benchmarks, spanning multiple backbone model sizes, demonstrate that AgenticRag-R1 consistently outperforms strong baselines. Moreover, AgenticRag-R1 learns more robust, interpretable, and
memory-aware
reasoning behaviors, highlighting the effect of fine-grained action modeling and information-aware optimization for long-horizon
reasoning. Our code is anonymous available at https://github.com/jiangxinke/Harness-RL/tree/AgenticRAG-R1-Whitebox.
Read the full abstract ▾
All authors Xinke Jiang, Yue Fang, Zhibang Yang, Jiaran Gao, Zhixin Zhang, Tao Feng, Rihong Qiu, Wentao Zhang, Hongxin Ding, Ruizhe Zhang, Yongxin Xu, Yuheng Huang, Xu Chu, Junfeng Zhao, Yasha Wang
Computer Vision
CineForge: Self-Improving Agents for Long-Horizon Video Generation
Junxiang Liu, Lin Wang, Haiyu Shi et al. (+10 more)
Long-horizon story-driven
video generation requires a production
agent to coordinate narrative decomposition, state tracking, shot design,
prompt construction, rendering, and revision across interdependent scenes. Existing adaptive
video systems primarily refine requests or reusable skills, leaving recurring production failures disconnected from persistent, stage-targeted improvements across stories. We introduce CineForge, a self-evolving
video-production
agent framework that couples CineForge-Produce for
video generation with CineForge-Evolve for cross-story policy
evolution. CineForge-Produce organizes each source story into typed narrative, character, spatial, and cinematic states, uses them to coordinate asset and clip generation, and records the process as a canonical production trajectory. CineForge-Evolve applies Case-to-Pattern-to-Policy
Evolution (CPPE) to review trajectory evidence, consolidate recurrent findings into bounded stage-local patches, and deploy validated updates through structural replay and confidence-controlled paired evaluation. To measure complete story realization, we introduce CineScope, which combines a 100-script CineScope-Data suite with a human-aligned, multiscale CineScope-Metric spanning causal state, directorial orchestration, pacing and resource allocation, and character arc. Across CineScope-Data and two public
benchmarks, the evolved CineForge policy improves CineScope-Metric from 4.024 to 4.380, outperforms three long-
video baselines with consistent gains under ScriptAgent, and reduces review LLM calls by 37.0% on new stories. These results establish production trajectories as actionable experience for
video agents that improve cumulatively across long-form storytelling tasks.
Read the full abstract ▾
All authors Junxiang Liu, Lin Wang, Haiyu Shi, Hongxu Ma, Xiaoyu Yang, Chunjie Chen, Xiaoxiao Xu, Kaiqiao Zhan, Boao Wang, Shuizhou Shi, Tianyun Zhu, Jie Li, Jiangtong Li
Information Theory
Multi-Access Speculative Inference: Uplink or Downlink?
Chang Cai and Kaibin Huang
Multi-access speculative inference (Multi-
SPIN) extends
SPIN to multi-device edge networks to accelerate cooperative
token generation. It allows on-device small
language models (SLMs) to autoregressively draft multiple
tokens for individual generation tasks, while an edge-server large
language model (LLM) verifies them in parallel. The major communication overhead arises when a drafted
token is rejected by the server, in which case sampling the correction
token requires access to both the SLM-output draft distribution and the LLM-output target distribution over the full
token vocabulary. Existing designs typically perform correction at the server by uploading the draft distribution, but transmitting a vocabulary-wide distribution creates a critical uplink (UL) bottleneck. Alternatively, the correction can be performed at the device by downloading the target distribution, leveraging the high transmission rates available on the downlink (DL). Motivated by this insight, we introduce communication-mode selection as a new design dimension for Multi-
SPIN. Specifically, each device can adaptively switch between the UL and DL modes to balance the UL bottleneck against the shared DL resource constraint, thereby relieving the overall communication burden. We formulate a sum-
token-goodput maximization problem that jointly accounts for mode selection, draft-length control, and power allocation. For mode selection, we reveal a simple optimal structure that enables efficient search over the number of UL devices, with the corresponding transmit powers optimized accordingly. For draft-length control, we develop a greedy-search
algorithm that adapts device-specific draft lengths to heterogeneous computation and communication capabilities. Experimental results on Qwen2.5 and DeepSeek-R1 model pairs demonstrate that the proposed framework significantly improves
token goodput.
Read the full abstract ▾
Computation and Language
Memory-First Fact-Checking: A Knowledge-Graph-Grounded Multi-Agent System for Misinformation Detection
Amelia Petrenciuc, Alexandru Lecu and Adrian Groza
This paper introduces a hybrid fact-checking framework that integrates Knowledge Graph-based semantic
memory with
adversarial multi-
agent reasoning for explainable misinformation detection. The proposed system follows a
memory-first, web-fallback architecture, in which input claims are initially evaluated against a dual-index Knowledge Graph through Sentence-BERT-based semantic
retrieval and Natural
Language Inference. When the evidence retrieved from the graph is insufficient to support a reliable decision, the framework collects information from trusted web sources and assesses it using an
adversarial tribunal composed of support, contradiction, and judging
agents. A graph-aware confidence mechanism combines semantic similarity, NLI confidence, and structural graph evidence to determine whether internal knowledge is sufficient, thereby reducing unnecessary web
retrieval. Following verification, validated information is transformed into structured triples and incorporated into the Knowledge Graph, supporting the incremental expansion of the system's semantic
memory. Experimental evaluation on a curated COVID-19 misinformation
benchmark demonstrates that the proposed framework achieves an accuracy of 97.4% and a macro-averaged F1-score of 92.6% on resolved claims, outperforming a Llama~3.3~70B baseline, which obtains an accuracy of 87.7% and a macro-averaged F1-score of 86.3%.
Read the full abstract ▾
Computation and Language
JPO: Juris Policy Optimization for Structured Legal Reasoning in Criminal Judgment Prediction
Zhaolu Kang, Yantao Liu, Tailong Luo et al. (+14 more)
Criminal judgment prediction requires models to infer statutory articles, charges, and sentencing outcomes from case facts. Unlike standard classification tasks, it involves a structured
reasoning process in which statutes should be matched with facts, charges should be justified by statutes, and sentencing outcomes should remain consistent with charges. Existing approaches optimize final labels, and while some have attempted to evaluate
reasoning quality, their evaluations are indirect, often relying on LLM-generated rubrics that reflect model-internal preferences rather than the inherent logical structure of legal adjudication. We propose Juris Policy Optimization (JPO), a post-training framework for structured legal
reasoning in Chinese criminal judgment prediction. JPO first uses teacher-generated rationales to supervise a standardized four-step
reasoning process, and then applies reinforcement learning with a composite reward over legal prediction quality,
reasoning structure completeness, and cross-step consistency. JPO further introduces
token-level advantage reweighting and adaptive clipping for legally salient
reasoning segments. Experiments on multiple open-source
language models and three Chinese legal
benchmarks show that JPO consistently improves both judgment prediction and
reasoning quality over supervised fine-tuning and reinforcement learning baselines.
Read the full abstract ▾
All authors Zhaolu Kang, Yantao Liu, Tailong Luo, Leqi Zheng, Lei Wei, Chenghua Zhu, Junhao Gong, Jiachen Qian, Eric Hanchen Jiang, Jiaxin Liu, Yuan Wang, Hao Zhang, Zixia Wang, Rong Fu, Zheng Lin, Richeng Xuan, Zhichao Hu
Multiagent Systems
Forward-Deployed Full-Stack Engineering for Autonomous Cloud MLOps
Sagar Srinivas Sakhinana and Venkataramana Runkana
Across industries, machine-learning systems support applications ranging from prediction and anomaly detection to
forecasting, optimization, and scheduling, yet operationalizing these systems requires coordinating application development, model pipelines, cloud infrastructure, security, deployment, monitoring, retraining, recovery, and rollback. We present an evidence-gated multi-
agent framework for transforming a natural-
language MLOps cloud engineering task into a verified repository and operational cloud deployment. The framework combines graph engineering, loop engineering, and
agent harness engineering. A stateful Graph Orchestrator coordinates specialized
agents for repository generation, review, execution, verification, release, and monitoring while governing workflow dependencies, evidence gates, retry bounds, recovery paths, and termination. Consequential lifecycle transitions proceed only when their required predicates are supported by verifiable execution or runtime evidence. Verification failures activate bounded reflection, repair, and re-verification, while runtime evidence of failure, drift, degradation, or policy violation can trigger bounded adaptation, recovery, or rollback.
Agent harness engineering constrains repository generation, review, and repair, artifact execution, and cloud operations through controlled capabilities and isolated execution environments. We realize the framework on Google Cloud Platform and evaluate repository completeness, controlled execution, evidence-gated transitions, cloud promotion, and bounded recovery. Our experimental results show that the framework prevents unsupported lifecycle transitions and drives each run toward either a verified operational deployment or an auditable terminal failure.
Read the full abstract ▾
Computation and Language
Cross-lingual Functional Vectors for Emotion Detection in Large Language Models
Jieying Xue, Phuong Minh Nguyen, Minh Le Nguyen et al. (+1 more)
Function vectors (FVs) have recently emerged as a promising mechanism for steering the behavior of large
language models (LLMs) by injecting task-specific latent direction representations derived from in-context demonstrations. While prior studies have shown that FVs can recover task behavior in structured in-context learning settings, their effectiveness on semantically complex tasks and their ability to generalize across
languages remain underexplored. We investigate the cross-lingual transferability of FVs using
multilingual multi-label
emotion recognition as a challenging semantic classification
benchmark. Specifically, we examine whether FVs extracted from a source
language can steer task behavior in another
language under both standard clean and perturbed zero-shot settings without providing demonstrations during inference. Across diverse cross-lingual settings, applying FVs substantially improves performance, suggesting that FVs capture
language-agnostic, task-relevant signals rather than purely
language-specific lexical patterns, and highlighting their potential as a lightweight and transferable mechanism for
multilingual task adaptation. We observe that each LLM exhibits a relatively stable optimal range of
attention heads for constructing effective FVs, and the pattern remains consistent across
languages. In addition, FVs can partially replicate the task-steering effects of standard few-shot in-context learning while avoiding the computational overhead of processing multiple demonstrations, making them effective for large-scale practical applications. Our code is available at https://github.com/yingjie7/cross_lingual_fvs.
Read the full abstract ▾
All authors Jieying Xue, Phuong Minh Nguyen, Minh Le Nguyen, Shogo Okada
Artificial Intelligence
LLMs Interpret, Embeddings Organize, Graphs Emerge: Agent-Driven Compilation of Scientific Knowledge
Shi-Ju Ran, Kun Zhang, Xi Wu et al. (+2 more)
Sustained scientific work requires a knowledge substrate that carries interpretation across tasks and preserves paths to source evidence. We call this process scientific knowledge compilation and implement it in ASKS, the
Agent-Driven Scientific Knowledge System. For each source, an LLM produces a readable Wiki view and machine-facing semantics. Deterministic checks convert the latter into a document-local GraphDelta, and
embedding geometry together with explicit graph rules integrates the proposed changes into persistent state. Each ingest is an inspectable state transition over accumulated knowledge, with compiled Wiki and graph views linked to the preserved source record. We examine this process by chronologically compiling 56 published papers from one research program. Branch survival, cross-paper support, lineage, coverage, and churn yield a source-traceable author research portrait centered on tensor-network methods, with branches into
quantum many-body research, tensor-network machine learning, and
quantum-AI-oriented directions. In this run, higher-level Hub organization remains stable and low-churn. Canonical-node growth is predominantly additive. Graph-level measurements and
navigation paths retain links to the source records from which they were compiled.
Read the full abstract ▾
All authors Shi-Ju Ran, Kun Zhang, Xi Wu, Liu-Si Yang, Wen-Jun Li
Computer Vision
See the Change, Keep the Flow: Unsupervised Action Segmentation via Spectral-Temporal Representation Learning
Yun Li, Jun Xiao, Cong Zhang et al. (+1 more)
Unsupervised action
segmentation aims to discover latent action categories and their temporal organization without action annotations. Optimal transport-based methods provide structured frame-to-action assignments, however, their pseudo-label quality is fundamentally conditioned on the representation space used to construct the transport cost. We argue that reliable OT pseudo-labeling requires a representation geometry that is simultaneously sensitive to discriminative action changes and coherent along local temporal progressions. Based on this insight, we propose SpecT-OT, a spectral-temporal representation learning framework built upon an unbalanced optimal transport pseudo-labeling concept. SpecT-OT introduces a Spectral Reparameterization Projector (SRP), which parameterizes projector weights with fixed
Fourier bases and learnable coefficients to improve the modeling of rapidly varying discriminative features, and Temporal Affinity Regularization (TAR), which imposes distance-aware, label-free constraints on pairwise frame affinities to stabilize local temporal structure. The two components jointly produce more discriminative and temporally stable transport costs, yielding more reliable pseudo-labels for iterative representation learning. Experiments on four
benchmarks demonstrate strong performance compared with state-of-the-art methods. SpecT-OT achieves the best results on 13 of 15 metrics, including 4.1-point MoF and 7.4-point F1 gains over the baseline on Breakfast and Desktop Assembly, respectively.
Read the full abstract ▾
All authors Yun Li, Jun Xiao, Cong Zhang, Kin-Man Lam
Computation and Language
Beyond Surface Alignment: Grounding the Dynamics of Situational Understanding and Generative Control in LLMs
Chenghao Yang
The current
alignment tuning paradigm for Large
Language Models (LLMs) prioritizes surface-level behaviors – fluency, safety, and tonal consistency. While effective for casual chat, this thesis argues that such surface
alignment masks a lack of grounding, creating models that are stylistically confident but situationally brittle. We propose a framework of Grounded
Alignment, analyzing how models process context (Input) and structure generation (Output), then aligning these grounded behaviors to human needs. First, we evaluate failures in Situational Grounding. SitTest shows that despite large context windows, state-of-the-art models struggle to maintain a consistent "mental model" of a changing environment. ReCode further shows that models rely on surface heuristics rather than deep syntactic dependencies: they "read" extensive histories without truly "understanding" the evolving situation. Second, we evaluate Generative Grounding. We introduce the Branching Factor (BF) to map LLM generation, finding that standard
alignment tuning constricts this landscape into premature stylistic collapse. Hindsight further shows that models often fail to understand their own generations. Finally, we propose Dynamic Control for grounded interaction. AI Realtor demonstrates context engineering to compensate for poor situational grounding. Base-Aligned Model Collaboration decouples exploration from stylistic constraints. We also present Annealed Sampling for verifiable reinforcement learning and apply these ideas to Addiction Support, where model-generated rationalization offers a communication interface for high-stakes domains. Collectively, this work moves beyond surface
alignment toward
agents anchored in both context and generation.
Read the full abstract ▾
Computer Vision
nnMNet: Baseline for Martian Terrain Semantic Segmentation
Ming-Han Lee and Chi-Yeh Chen
Semantic
segmentation is a crucial task for understanding Mars, the most Earth-like
planet in our
solar system. However, it is challenging because the Martian surface is highly unstructured and complex, making accurate pixel-level prediction and fine-grained annotation difficult. Recent advancements in deep learning have introduced numerous methods and
datasets to address these challenges. Nevertheless, the field lacks a robust, publicly available, and reproducible baseline, as well as a unified
benchmark to facilitate fair evaluations. In this work, we present nnMNet, a new baseline model designed for Martian terrain semantic
segmentation. Building upon nnWNet, we integrate linear
attention to better capture global context and employ lightweight convolutions to reduce computational overhead. To bridge the gap between local and global representations, we introduce the Spatially-Aware Fusion Block (SAFB), which augments and combines features with diverse characteristics. Furthermore, we establish a new
benchmark by curating and standardizing three high-quality
datasets for thorough evaluation. nnMNet achieves new state-of-the-art 86.61%, 83.25%, and 88.24% mIoU on SynMars-TW, SynMars-Air, and MarsScapes, respectively. Our code, models, and
datasets are publicly available at https://github.com/dereklee0310/nnMNet.
Read the full abstract ▾
Machine Learning
Wide Learning: Learning to Reach Evidence
Junzhou Chen
Machine learning is usually evaluated after an evidence interface has been fixed. A
dataset, sensor suite, query
language, action set, or experimental protocol determines which observations can be obtained, and learning is judged by what it extracts from them. We study a complementary capability. A learner's state can determine which evidence-generating experiments it can reliably realise under bounded resources, even when primitive affordances remain fixed. We call this learner-relative experiment family its effective epistemic reach, and use Wide Learning for task-relevant learning-induced changes in that family.We formalise effective reach relative to learner state, deployment budget, reliability threshold, and evaluation distribution. In a controlled construction, two hidden worlds have exactly the same public observation law. An informative diagnostic exists in a fixed five-primitive substrate. Before calibration, one address attempt realises it with probability at most $2^{-10} = 1/1024$, below a pre-specified 0.95 threshold; after calibration, held-out realisation is 1. Public-channel total variation is 0, whereas the realised diagnostic has total variation 1, and sealed binary risk moves from approximately 1/2 to 0. The construction establishes that learning can change effective epistemic reach even when primitive affordances and deployment resources are held fixed. It opens a complementary evaluation question for learning systems: not only what they infer from available evidence, but what informative evidence experience teaches them to bring within reach.
Read the full abstract ▾