Search arXiv⌕ Search

arXiv subjects

Dawei Yin

Publications and source records attributed to Dawei Yin.

At least 19 recordsLinked to original sources
arXiv2026

FDE-Bench: Evaluating End-to-End Delivery from Underspecified Real-World Business Requests

Real-world business requests often leave key requirements unstated because frontline practitioners struggle to articulate rules they know from experience. Forward-deployed engineers (FDEs) must elicit these requirements through dialogue and iteratively optimize solutions until they meet customer acceptance criteria. Existing optimization benchmarks typically start from explicit objectives, while clarification benchmarks often introduce information gaps into existing tasks. We introduce FDE-Bench, the first benchmark built from real customer projects to evaluate agents on end-to-end delivery tasks that require interactive requirement clarification and iterative optimization. Its 49 cases span combinatorial optimization and machine learning. Clarification targets are distilled from discussions between frontline practitioners and FDEs. Agents elicit missing requirements from a simulated customer, and case-specific evaluators assess final artifacts against acceptance criteria using customer-accepted FDE solutions as quality references. Across 18 agent configurations, the highest mean delivery score is only 0.5510, compared with an FDE reference of 1.0. The highest delivery success rate (DSR), defined as the fraction of planned runs reaching this reference, is only 6.12%. These results show that reliable FDE-level end-to-end delivery remains out of reach for the evaluated agents on underspecified real-world business tasks. The benchmark project page is available at https://baidubce.github.io/FM-Agent/FDE-Bench/.

cs.AI↗
arXiv2026

AGAR: a reinforcement learning substrate for LLM program evolution

Given a task and an evaluator, a language model can rewrite a candidate program while a search loop decides which rewrites survive, offering a practical route to algorithm discovery. But that loop is governed by five constants set by hand: which parent to select, how hard to mutate, how to keep diversity, what to remember, and a scalar score that never says which part of the program earned it. Reinforcement learning already has an estimator for each. The obstacle is that program evolution is not usually written down as a decision process. We formalize it as a Markov decision process whose action is the modular prefix the model is conditioned on, rather than the program it emits. Credit assignment, value estimation, adaptive exploration, and experience memory can then attach to distinct components. AGAR (Algorithm Generation As RL) provides the resulting substrate: any estimator can be replaced or switched off without changing the controller, making the transfer auditable one mechanism at a time, with no gradient training of the backend model. Across 19 tasks, two backends, and three seeds under one harness, AGAR improves on the stronger of two published baselines on most tasks, with gains concentrated in the competitive-programming family. The formalization also yields a checkable reading of prior work: these systems are implicitly zero-discount, not by choice, but because fitness is exogenous to an individual rather than a return over successors, leaving a discount factor nothing to act on.

cs.AI↗
arXiv2026

RLDISCOVER: LLM-driven co-evolution of reinforcement learning algorithms

LLM-guided program evolution has enabled discoveries in mathematics and computational optimization, raising the prospect of reinforcement learning (RL) algorithms that self-evolve to improve how agents learn. However, realizing this prospect faces two obstacles. Joint search over coupled algorithmic components is difficult to scale: simultaneous changes can disrupt learning, while isolated changes overlook their dependencies. Evaluating candidate algorithms also requires costly training, with fitness remaining uncertain across random seeds. We introduce RLDiscover, a framework for the self-evolution of model-free deep RL algorithms. Progressive Co-Evolution advances from targeted component edits to joint evolution, while Progressive Probabilistic Evaluation balances search breadth and evaluation fidelity through staged training and repeated evaluation. Experiments across SAC, PPO, and DQN on four benchmark suites show substantial improvements in mean return, with per-family median gains of 32%-84% and a peak return ratio of approximately 363x over a near-zero baseline. These gains include transitions from failed learning to successful task completion, and improvements persist when evolution starts from stronger open-source implementations. On measured SAC locomotion runs, evaluation uses approximately one-fifteenth the estimated compute required to fully evaluate the same candidate pool. Remarkably, independent searches repeatedly discover interpretable combinations of adaptive robust losses, progress-dependent value targets, and running statistics, with selected programs transferring to unseen tasks. These findings point toward a broader role for self-evolution in AI: discovering interpretable algorithms that improve how agents learn.

cs.AI↗
arXiv2026

QAM: Quadratic-Accurate Checkpoint Merging via Sequential Consistency

Saved checkpoints record states along a training trajectory, but generally do not determine the updates at states that would be visited under a different schedule. We study how accurately these checkpoints can reconstruct the endpoint of a sequential reference with prescribed update strengths. Under a common local transition model, two checkpoint-index moment conditions characterize all convex merges that agree with this reference through second order. We then prove an information limit that for nondegenerate profiles, no algorithm using only a fixed-length gradient-descent (GD) history with step size $h$ can achieve $o(h^3)$ endpoint error uniformly over a fixed class of smooth, strongly convex losses. The lower bound follows from two losses with identical GD checkpoint histories but sequential reference endpoints separated by $Ω(h^3)$. \textbf{Quadratic-Accurate Merging} (QAM) achieves a matching uniform $O(h^3)$ endpoint error bound. Its explicit coefficients also define the unique profile-dependent merge that exactly matches the sequential GD reference across all fixed quadratic objectives. Across two public Adam checkpoint trajectories (SmolLM3-3B and OpenEuroLLM-Prelude-9B), three windows and three profiles per model, and 15 tasks, QAM shows mixed results for short windows and broader advantages over \textbf{Warmup-Stable and Merge} (WSM) for longer windows. Matched-moment GSM8K diagnostics further show that local consistency alone does not fully determine downstream scores. These results characterize the reconstruction limits of saved histories, provide a coefficient rule that attains the optimal rate, and assess its practical utility.

cs.LG↗
arXiv2026

From Transient Prompts to Persistent Control: Scientific Poster Generation via Recursive Semantic-Geometric Contracts

Scientific poster generation distills a multimodal paper into a single-page visual artifact, forcing strict trade-offs between informational coverage and readability under a fixed spatial budget. Existing methods pass plans as transient prompts and validate individual stages in isolation. This strategy causes requirements to drift across content and layout modules, and previous checks to be silently invalidated. We introduce PosterVisor, a control framework that shifts poster generation from transient prompts to persistent control. An Orchestrator grounds rubrics in the paper and visual assets, compiling them into a Semantic-Geometric Contract (SGC) that binds claims and sources to required visuals, budgets, and spatial commitments. Only fully instantiated records become executable assertions; other usable requirements remain soft guidance. Recursive Contract Enforcement (RCE) dynamically triggers checks across stages as evidence emerges. Crucially, during repairs, RCE rechecks affected checkpoint states, preventing repair-induced regressions from propagating silently. We instantiate PosterVisor in HTML/CSS and editable PPTX generators. On the 100-paper Paper2Poster benchmark, PosterVisor-PPT improves observed mean poster-grounded QA accuracy over PosterGen (64.47% vs. 58.53%) and is preferred by human judges in 72.5% of non-tied pairwise comparisons (95% CI, 61.6-83.4%). A secondary 30-paper study also yields higher VLM Overall and PaperQuiz means. These results support rubric-compiled contracts and stage-conditioned enforcement for controllable poster synthesis.

cs.AI↗
arXiv2026

Navigating Sparse Evidence: Agentic Visual RAG via Explicit Context Selection and Consolidation

Visual Retrieval-Augmented Generation (VRAG) empowers models to navigate and answer queries about visually rich documents by retrieving relevant page images as visual evidence and reasoning over their content. However, effectively utilizing this visual evidence is usually impeded by two main challenges. First, answer-relevant evidence is sparse and may be concentrated in a small region of one page or dispersed across multiple pages. Second, existing agentic methods often generate answers based on raw exploration trajectories or compressed textual memories rather than an explicitly organized set of supporting images, making answers susceptible to exploration noise and obscuring the evidence-backed reasoning trace. We argue that the bottleneck lies not only in evidence discovery but also in its preservation and organization before answer generation. We propose SCoRE (Selection and Consolidation for Robust Evidence), a unified agent loop for explicit evidence selection and consolidation. During exploration, SCoRE retains only query-relevant observations and their source pointers in a maintained textual ledger, preserving earlier evidence while keeping the visual context bounded. At termination, it reloads the referenced original images and consolidates the visual evidence for answering, arranging it into a logical sequence. This decouples final reasoning from exploratory trial-and-error while ensuring strict visual grounding via indexed claim-to-image linkages. To enable end-to-end optimization of this unified rollout, our training paradigm combines filtered cold-start trajectory distillation with evidence-aware reinforcement learning, whose reward promotes evidence coverage, consolidation compactness, and answer correctness.

cs.AI↗
arXiv2026

SWRouter: Similarity-Contractive Window Routing for Multi-Turn Large Language Model Conversations

Large language models exhibit complementary strengths, motivating routing methods that dispatch each query to the most suitable model. Although existing routers are effective in single-turn settings, they do not directly transfer to multi-turn dialogue, where routing performance critically depends on how historical context is segmented, retained, and incorporated into the current prompt. This introduces two fundamental challenges: preventing information loss and information confusion during context construction, and evaluating routing quality without conflating model selection with prompt construction quality. In this paper, we propose SWRouter, a Similarity-Contractive Window Router for multi-turn large language model routing. SWRouter combines a similarity-based context segmentation mechanism for prompt construction with a dual-metric evaluation framework that decouples construction accuracy from router performance. Experiments on multi-turn dialogue benchmarks demonstrate that SWRouter consistently surpasses strong baselines, achieving a 16.26% improvement in evaluation accuracy over the best individual large language model and an additional 8.22% gain over the Conv-ID Context baseline. Our results highlight that multi-turn large language model routing requires a joint design of context construction and evaluation, rather than a direct extension of single-turn routing methods.

cs.CL↗
arXiv2026

EndPrompt: Efficient Long-Context Extension via Terminal Anchoring

Extending the context window of large language models typically requires training on sequences at the target length, incurring quadratic memory and computational costs that make long-context adaptation expensive and difficult to reproduce. We propose EndPrompt, a method that achieves effective context extension using only short training sequences. The core insight is that exposing a model to long-range relative positional distances does not require constructing full-length inputs: we preserve the original short context as an intact first segment and append a brief terminal prompt as a second segment, assigning it positional indices near the target context length. This two-segment construction introduces both local and long-range relative distances within a short physical sequence while maintaining the semantic continuity of the training text--a property absent in chunk-based simulation approaches that split contiguous context. We provide a theoretical analysis grounded in Rotary Position Embedding and the Bernstein inequality, showing that position interpolation induces a rigorous smoothness constraint over the attention function, with shared Transformer parameters further suppressing unstable extrapolation to unobserved intermediate distances. Applied to LLaMA-family models extending the context window from 8K to 64K, EndPrompt achieves an average RULER score of 76.03 and the highest average on LongBench, surpassing LCEG (72.24), LongLoRA (72.95), and full-length fine-tuning (69.23) while requiring substantially less computation. These results demonstrate that long-context generalization can be induced from sparse positional supervision, challenging the prevailing assumption that dense long-sequence training is necessary for reliable context-window extension. The code is available at https://github.com/clx1415926/EndPrompt.

cs.CL↗
arXiv2026

Graph Foundation Models for Recommendation: A Comprehensive Survey

Recommender systems (RS) serve as a fundamental tool for navigating the vast expanse of online information, with deep learning advancements playing an increasingly important role in improving ranking accuracy. Among these, graph neural networks (GNNs) excel at extracting higher-order structural information, while large language models (LLMs) are designed to process and comprehend natural language, making both approaches highly effective and widely adopted. Recent research has focused on graph foundation models (GFMs), which integrate the strengths of GNNs and LLMs to model complex RS problems more efficiently by leveraging the graph-based structure of user-item relationships alongside textual understanding. In this survey, we provide a comprehensive overview of GFM-based RS technologies by introducing a clear taxonomy of current approaches, diving into methodological details, and highlighting key challenges and future directions. By synthesizing recent advancements, we aim to offer valuable insights into the evolving landscape of GFM-based recommender systems.

cs.IR↗
arXiv2026

Not All Preferences Deserve Gradients: Understanding Gradient Utility in Offline Reasoning Alignment

Offline preference optimization aligns reasoning models from fixed chosen--rejected pairs, yet standard methods apply gradient updates from every pair regardless of its training value under the current policy. We argue that this uniform treatment is wasteful and potentially harmful. From the perspective of gradient utility, we show that a pair's contribution depends jointly on informativeness and stability. Pair utility drifts as the policy evolves, high-gradient samples can coincide with high-curvature regions, leading to noisy and destabilizing updates, and the most effective supervision comes from stable confident errors where the model is reliably wrong yet curvature remains low. These findings motivate SAGE (Stability-Aware Gradient Efficiency), which maintains difficulty-stratified candidate pools refreshed during training and selects pairs within each pool by a forward-pass signal-to-curvature score. Only pairs with high current utility receive gradient computation; the rest are excluded from backpropagation. On mathematical reasoning benchmarks across multiple model scales, SAGE outperforms full-data and size-matched baselines while producing substantially smoother optimization trajectories.

cs.AI↗
arXiv2026

Learning a Single Token to Replace Long System Prompts in LLMs

Long system prompts are widely used to steer Large Language Models (LLMs), but repeatedly processing them at inference time is inefficient and consumes valuable context budget. This motivates a central question: can the behavioral effect of a long system prompt be retained using only a minimal learned representation? To enable this, we propose a lightweight training framework that learns a single Behavior-Equivalent Token ([BE]). The framework first trains [BE] to encode the semantic content of the original system prompt via reconstruction, and then distills the prompt's downstream behavior into this single token. Importantly, our method requires no update to the pretrained LLM weights, no auxiliary compression models, and no labeled responses. Empirical evaluations on three datasets show that replacing long prompts with a single [BE] token yields up to a $3000\times$ prompt compression ratio, while retaining about 98% of the downstream performance of the original system prompts. This substantially reduces inference cost and frees nearly the entire context window for user inputs and model outputs.

cs.CL↗
arXiv2026

Unleashing the Power of LLMs in Dense Retrieval with Query Likelihood Modeling

Dense retrieval is a crucial task in Information Retrieval (IR), serving as the basis for downstream tasks such as re-ranking and augmenting generation. Recently, large language models (LLMs) have demonstrated impressive semantic understanding capabilities, making them attractive to researchers focusing on dense retrieval. While LLMs, as decoder-style generative models, excel in language generation, they often fall short in modeling global information due to a lack of attention to subsequent tokens. Drawing inspiration from the classical word-based language modeling approach for IR, specifically the query likelihood (QL) model, we aim to leverage the generative strengths of LLMs through QL maximization. Rather than employing QL estimation for document ranking, we propose an auxiliary task of QL maximization to enhance the backbone for subsequent contrastive learning of the retriever. We introduce our model, LLM-QL, which incorporates two key components: Attention Block (AB) and Document Corruption (DC). AB blocks the attention of predictive tokens to the document tokens before the document's ending token, while DC corrupts a document by masking a portion of its tokens during prediction. Evaluations on the in-domain (MS MARCO) and out-of-domain dataset (BEIR) indicate LLM-QL's superiority over other LLM-based retrievers. Furthermore, comprehensive analyses also validate the efficacy of LLM-QL and its components.

cs.IR↗
arXiv2026

LLM-Specific Utility for Retrieval-Augmented Generation

Retrieval-augmented generation (RAG) is typically optimized for topical relevance, yet its success ultimately depends on whether retrieved passages are useful for a large language model (LLM) to generate correct and complete answers. We argue that such utility is often LLM-specific rather than universal, due to differences in models' knowledge, reasoning, and ability to leverage evidence. We formalize LLM-specific utility as the performance improvement of a target LLM when a passage is provided, compared to answering without evidence. To systematically study LLM-specific utility, we construct a benchmark of LLM-specific gold utilitarian passages for four LLMs (Qwen3-8B/14B/32B and Llama 3.1-8B) on three QA datasets (Natural Questions, TriviaQA, and MS MARCO-FQA). Our analysis shows that utilitarian passages are model-dependent and non-transferable: each LLM performs best with its own utilitarian evidence, while evidence optimized for other LLMs is consistently suboptimal. Human-annotated evidence remains a strong general baseline but does not fully match individual LLM utility needs. We further introduce the LLM-specific utility judgment task and construct the corresponding benchmark, i.e., SpecUBench (LLM-Specific Utility Benchmark). Experiments show that existing utility-aware selection and scoring methods largely capture model-agnostic usefulness and struggle to reliably estimate LLM-specific utility. Overall, our findings highlight the limitations of current utility-aware retrieval and motivate generator-tailored evidence selection for improving RAG. Our code and datasets can be found at https://github.com/Trustworthy-Information-Access/LLM_specific_utility.

cs.CL↗
arXiv2026

DuMateBench: Evaluating Autonomous Agents in Complex Real-World Workflows

Autonomous agents are increasingly adopted to complete complex, multi-tool workflows in real-world settings. However, existing benchmarks typically separate tasks by application or capability and evaluate agents in environments that are cleaner and more stable than those encountered in practice. We introduce DuMateBench, a real-session benchmark reconstructed from anonymized and privacy-screened user sessions collected from a large-scale production agent platform. Each task preserves the relevant pre-solution interaction history, persistent configurations, and workspace state, and is then validated through human verification. The resulting benchmark comprises 200 tasks spanning 8 broad scenarios and 17 fine-grained capability categories, with most tasks requiring multiple capability coordination. We execute these tasks in isolated Docker containers injected with three forms of real-world environmental complexity: Insufficient, Unstable, and Noisy, and assess performance using a hybrid deterministic and LLM-as-Judge evaluation protocol. Experiments across five representative autonomous-agent frameworks paired with four state-of-the-art LLMs reveal substantial gaps in strict task completion. Complementary robustness, efficiency, and diagnostic analyses further show that performance under environmental perturbations is jointly shaped by the capabilities of the LLM and the surrounding agent framework. The code and data are publicly available at https://dumatebench.com/.

cs.AI↗
arXiv2026

Can LLM Annotations Replace User Clicks for Learning to Rank?

Large-scale supervised data is essential for training modern ranking models, but obtaining high-quality human annotations is costly. Click data has been widely used as a low-cost alternative, and with recent advances in large language models (LLMs), LLM-based relevance annotation has emerged as another promising annotation. This paper investigates whether LLM annotations can replace click data for learning to rank (LTR) by conducting a comprehensive comparison across multiple dimensions. Experiments on both a public dataset, TianGong-ST, and an industrial dataset, Baidu-Click, show that click-supervised models perform better on high-frequency queries, while LLM annotation-supervised models are more effective on medium- and low-frequency queries. Further analysis shows that click-supervised models are better at capturing document-level signals such as authority or quality, while LLM annotation-supervised models are more effective at modeling semantic matching between queries and documents and at distinguishing relevant from non-relevant documents. Motivated by these observations, we explore two training strategies -- data scheduling and frequency-aware multi-objective learning -- that integrate both supervision signals. Both approaches enhance ranking performance across queries at all frequency levels, with the latter being more effective. Our code is available at https://github.com/Trustworthy-Information-Access/LLMAnn_Click.

cs.IR↗
arXiv2026

VISOR: Agentic Visual Retrieval-Augmented Generation via Iterative Search and Over-horizon Reasoning

Visual Retrieval-Augmented Generation (VRAG) empowers Vision-Language Models to retrieve and reason over visually rich documents. To tackle complex queries requiring multi-step reasoning, agentic VRAG systems interleave reasoning with iterative retrieval. However, existing agentic VRAG faces two critical bottlenecks. (1) Visual Evidence Sparsity: key evidence is scattered across pages yet processed in isolation, hindering cross-page reasoning; moreover, fine-grained intra-image evidence often requires precise visual actions, whose misuse degrades retrieval quality; (2) Search Drift in Long Horizons: the accumulation of visual tokens across retrieved pages dilutes context and causes cognitive overload, leading agents to deviate from their search objective. To address these challenges, we propose VISOR (Visual Retrieval-Augmented Generation via Iterative Search and Over-horizon Reasoning), a unified single-agent framework. VISOR features a structured Evidence Space for progressive cross-page reasoning, coupled with a Visual Action Evaluation and Correction mechanism to manage visual actions. Additionally, we introduce a Dynamic Trajectory with Sliding Window and Intent Injection to mitigate search drift. They anchor the evidence space while discarding earlier raw interactions, preventing context from being overwhelmed by visual tokens. We train VISOR using a Group Relative Policy Optimization-based Reinforcement Learning (GRPO-based RL) pipeline with state masking and credit assignment tailored for dynamic context reconstruction. Extensive experiments on ViDoSeek, SlideVQA, and MMLongBench demonstrate that VISOR achieves state-of-the-art performance while maintaining reasonable and controllable computational costs for long-horizon visual reasoning tasks. The source code is available at https://github.com/syc1336/VISOR.

cs.CV↗
arXiv2026

Janus: An Algorithm-Evaluator Co-Evolution Framework for LLM-Driven Discovery under Expensive Evaluation Budgets

LLM-driven program discovery relies on rapid evaluator feedback, but many scientific and engineering tasks require high-fidelity simulations, hardware execution, or physical experiments, making each evaluation expensive. Cheap surrogate evaluators can reduce this cost, yet fixed surrogates are vulnerable to search-induced distribution shift and are difficult to fit reliably from sparse, search-biased labels. We introduce Janus, a framework that uses LLMs to co-evolve target programs and executable proxy evaluators. To address label scarcity, Janus leverages domain knowledge encoded in LLMs to generate task-specific evaluator programs and calibrates them using real outcomes. To mitigate distribution shift, Janus evolves evaluators alongside target programs, selects them using a promotion-aligned objective, and maintains region-conditioned portfolios with online credit updates. Because proxy predictions remain fallible, Janus uses them only to prioritize candidates and requires real validation before candidates can enter the target-program population or update the incumbent. Across five scientific and engineering design tasks, Janus achieves a larger area under the best-so-far improvement curve over the real-evaluation budget and higher final performance than a matched baseline that evolves only target programs. On average, Janus reaches 99/% of the baseline's final improvement with 59.1/% fewer real evaluations. Evolved proxy evaluators also rank promising candidates more accurately than their seed versions. Together, these results extend evaluator-guided LLM discovery from tasks with cheap, scalable feedback to scientific domains where trustworthy evaluation is scarce and expensive.

cs.AI↗
arXiv2026

DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations

As autonomous agents rapidly evolve, their ability to reliably manipulate ubiquitous digital documents has become critical for enabling general-purpose AI assistants and automating complex workspace workflows. In this paper, we introduce DocOps, a deterministically verifiable evaluation framework underpinned by a hierarchical taxonomy that deconstructs document operations inspired by real-world practices into atomic dimensions and escalating workflow complexities. Based on DocOps, we systematically evaluate representative closed- and open-source models across various agentic harnesses, revealing that even the most advanced frontier configurations still exhibit profound limitations when handling highly coupled, long-range tasks. Furthermore, a fine-grained analysis of existing agents' manipulation behaviors uncovers 3 key failure modes: long-term state tracking collapse, shallow semantic verification, and destructive editing of structural metadata. Ultimately, our work exposes the capability boundaries of agents in maintaining global document consistency, shedding light on the future design of robust, non-destructive agents for complex digital ecosystems.

cs.AI↗