arXiv subjects

Wenjie Ma

Publications and source records attributed to Wenjie Ma.

At least 19 recordsLinked to original sources

arXiv2026

RAG over Thinking Traces Can Improve Reasoning Tasks

Retrieval-augmented generation (RAG) has proven effective for knowledge-intensive tasks, but is widely believed to offer limited benefit for reasoning-intensive problems such as math and code generation. We challenge this assumption by showing that the limitation lies not in RAG itself, but in the choice of corpus. Instead of retrieving documents, we propose retrieving thinking traces, i.e., intermediate thinking trajectories generated during problem solving attempts. We show that thinking traces are already a strong retrieval source, and further introduce T3, an offline method that transforms them into structured, retrieval-friendly representations, to improve usability. Using these traces as a corpus, a simple retrieve-then-generate pipeline consistently improves reasoning performance across strong models and benchmarks such as AIME 2025--2026, LiveCodeBench, and GPQA-Diamond, outperforming both non-RAG baselines and retrieval over standard web corpora. For instance, on AIME 2025-2026, RAG with traces generated by Gemini-2-thinking achieves relative gains of +56.3%, +8.6%, and +7.6% for Gemini-2.5-Flash, GPT-OSS-120B, and GPT-5, respectively, even though these are more recent models. Overall, our results suggest that thinking traces are an effective retrieval corpus for reasoning tasks, and transforming them into structured, compact, or diagnostic representations unlocks even stronger gains. Code available at https://github.com/Narabzad/t3.

cs.IR↗

arXiv2026

Quantum Structural Renormalization and Anharmonic Stabilization of Superconductivity in P4/mmm YScH8

The recent synthesis of P4/mmm-YScH$_{8}$ at 191.7 GPa with a superconducting critical temperature Tc of 113 K has attracted considerable interest in the study of ternary superhydrides. Here we systematically evaluate the effects of quantum and anharmonic motion of ions in P4/mmm-YScH$_{8}$ using the stochastic self-consistent harmonic approximation (SSCHA). We find that these effects renormalize the crystal structure and lower the dynamical stability threshold pressure from $\sim$140 GPa to $\sim$84 GPa, a reduction of about 40%. Furthermore, neglecting these effects causes a significant overestimation of the $T_{c}$, for instance by $\sim$20 K at 190 GPa. Our prediction of $T_{c}$ at 190 GPa is 113 K ($μ^{*}$= 0.13), close to the experimental observation. Analysis of the phonon density of states and Eliashberg spectral function further reveals site-dependent contributions to superconductivity. These results establish that a consistent treatment of nuclear quantum motion and anharmonicity is essential for accurately predicting superconductivity in compressed ternary hydrides.

cond-mat.supr-con↗

arXiv2026

Structural and Vibrational Properties of D$_3$Se from First Principles: Anharmonic Quantum and Isotope Effects

Hydrogen-rich superconductors have garnered considerable interest following the discovery of hot superconductivity in high pressure H$_3$S, reviving prospects for room temperature superconductors under high-pressures. Using H$_3$Se as a reference system, we investigate the vibrational and superconducting properties of D$_3$Se in the Im\bar{3}m phase across 60-200 GPa by combining first-principles calculations with the stochastic self-consistent harmonic approximation to treat ionic quantum and anharmonic effects. These effects introduce significant renormalization to the phonon spectra and stabilize the lattice down to at least 70 GPa, well below the harmonic prediction of >110 GPa. Ultimately, the phonon renormalizations alter the electron-phonon coupling, introducing a decrease in the superconducting critical temperature by about 3-16 K across the studied pressure range with respect to standard calculations. Including anharmonic phonons within the Migdal-Eliashberg theory yields $T_c \approx 154$ K at 75 GPa (with $μ^* = 0.1$, $λ = 3.0$), highlighting D3Se as a promising high-Tc superconductor at moderate pressures. Examining the role of anharmonicity in the isotope effect, we find that at 200 GPa it suppresses the isotope coefficient $α$ to 0.29 one third below the harmonic value (0.44) which approaches the BCS limit of 0.5. This dramatic reduction demonstrates that anharmonicity fundamentally governs the isotope effect on this system. The stark discrepancy between anharmonic and harmonic descriptions underscores the need for targeted experimental efforts to resolve the origin of the persistent theory-experiment discrepancy in compressed hydrides.

cond-mat.supr-con↗

arXiv2026

Euclid-Omni : A Unified Neuro-Symbolic Framework for Plane Geometry

Euclidean geometry is a compelling testbed for AI reasoning, as it demands the combination of intuitive diagram understanding, axiomatic deduction, and algebraic computation. Yet, existing approaches typically address only a subset of these abilities or struggle with competition-level problems. We introduce \textit{Euclid-Omni}, a unified neuro-symbolic framework that couples a formal geometry system with Large Language Models (LLMs) and Vision-Language Models (VLMs) to tackle both calculation- and proving-style problems, in formal and natural languages, up to Olympiad-level difficulty. At its core, we develop \textit{Euclidea}, a versatile symbolic geometry solver that automatically generates reasoning steps through deductive inference and algebraic computation. Building on this, we develop a data-generation pipeline that synthesizes symbolic problems and solutions, renders diagrams, and translates them into natural language, producing large-scale, diverse datasets for training LLMs and VLMs across a wide range of reasoning settings. Experiments show that VLMs trained on our synthetic data achieve superior performance on calculation tasks, and that LLMs combined with \textit{Euclidea} are competitive with state-of-the-art systems on Olympiad-level proving problems, despite using orders of magnitude less compute and training data. Code and scripts are publicly available at https://github.com/20171130/Euclid-Omni

cs.AI↗

arXiv2026

IEA: Amateur-Friendly Conversational Image Editing Agent via Three Stages of Multitask Alignment

Current image editing software often hinges on fixed filters or expert tuning, leaving a gap between amateur users' intent and outcomes. Creations by generative models may contain artifacts, implausible details, or stylistic drift away from photorealism and offer little insight into why an edit was made. We propose IEA, a conversational Image Editing Agent that learns to operate parameterized tools in an explicit, interpretable action space. IEA is trained via a three-stage multitask pipeline: (1) SFT on distilled expert edits, (2) GRPO with rewards for likeness improvement, tool usefulness, and intent summarization, and (3) large-scale synthetic fine-tuning to jointly master image editing, refinement, and user intent summarization. By manipulating 16 editing tools step by step, IEA produces transparent edit traces that can be inspected and debugged. In quantitative experiments, it attains a lower pixel distance on the edit task and a higher ROUGE-L on the summary task than strong baselines. In user studies, it ranks best among tool-calling methods for instruction following while surpassing generative methods in overall perceptual quality. Our results validate interpretable, tool-centric VLMs as a reliable path to human instruction-guided image retouching.

cs.CV↗

arXiv2026

BenchEvolver: Frontier Task Synthesis via Solution-Centric Evolution

The rapid progress of frontier large language models has led to widespread benchmark saturation, limiting the ability of existing datasets to differentiate model capabilities or provide useful training signal. For instance, on LiveCodeBench, frontier models achieve over 99% Pass@1 on easy splits and exceed 90% Pass@1 on average across difficulty levels. Constructing new, challenging datasets typically requires substantial human effort, creating a bottleneck for progress. We introduce BenchEvolver, a solution-centric evolutionary framework that automatically transforms existing coding problems into harder variants. Rather than generating problems from scratch, BenchEvolver evolves reference solutions through structured transformations and derives corresponding statements and tests from the evolved solutions. This design grounds generation in executable semantics, enabling scalable construction of high-quality, diverse, and difficult tasks with verifiable correctness. Applying BenchEvolver to LiveCodeBench and SciCode, we obtain evolved tasks that are substantially harder while maintaining validity, reference correctness, and diversity. We further curate LiveCodeBench-Plus, a 91-problem benchmark combining evolved and difficult original LCB-v6 tasks, where frontier-model Pass@1 ranges from 27.5% to 62.6%, restoring clear discrimination among strong coding models. Importantly, evolved tasks remain challenging even for the model that generates them, enabling self-improvement. We further show that RL on evolved LCB tasks improves held-out coding performance: for gpt-oss-20b, seed+evolved training achieves +8.7 and +8.3 Pass@1 gains on LCB v6 Hard and LCB-Pro Easy, exceeding seed-only gains by 70.7% and 34.8%, respectively. Our results show that BenchEvolver can convert saturated benchmarks into frontier-level evaluation suites and reusable training signal.

cs.SE↗

arXiv2026

optimize_anything: A Universal API for Optimizing any Text Parameter

Can a single LLM-based optimization system match specialized tools across fundamentally different domains? We show that when optimization problems are formulated as improving a text artifact evaluated by a scoring function, a single AI-based optimization system-supporting single-task search, multi-task search with cross-problem transfer, and generalization to unseen inputs-achieves state-of-the-art results across six diverse tasks. Our system discovers agent architectures that nearly triple Gemini Flash's ARC-AGI accuracy (32.5% to 89.5%), finds scheduling algorithms that cut cloud costs by 40%, generates CUDA kernels where 87% match or beat PyTorch, and outperforms AlphaEvolve's reported circle packing solution (n=26). Ablations across three domains reveal that actionable side information yields faster convergence and substantially higher final scores than score-only feedback, and that multi-task search outperforms independent optimization given equivalent per-problem budget through cross-task transfer, with benefits scaling with the number of related tasks. Together, we show for the first time that text optimization with LLM-based search is a general-purpose problem-solving paradigm, unifying tasks traditionally requiring domain-specific algorithms under a single framework. We open-source optimize\_anything with support for multiple backends as part of the GEPA project at https://github.com/gepa-ai/gepa .

cs.CL↗

arXiv2026

DreamProver: Evolving Transferable Lemma Libraries via a Wake-Sleep Theorem-Proving Agent

We introduce DreamProver, an agentic framework that leverages a "wake-sleep" program induction paradigm to discover reusable lemmas for formal theorem proving. Existing approaches either rely on fixed lemma libraries, which limit adaptability, or synthesize highly specific intermediate lemmas tailored to individual theorems, thereby lacking generality. DreamProver addresses this gap through an iterative two-stage process. In the wake stage, DreamProver attempts to prove theorems from a training set using the current lemma library while proposing new candidate lemmas. In the "sleep" stage, it abstracts, refines, and consolidates these candidates to compress and optimize the library. Through this alternating cycle, DreamProver progressively evolves a compact set of high-level, transferable lemmas that can be effectively used to prove unseen theorems in related domains. Experimental results demonstrate that DreamProver substantially improves proof success rates across a diverse set of mathematical benchmarks, while also producing more concise proofs and reducing computational cost.

cs.AI↗

arXiv2026

Reliable Fine-Grained Evaluation of Natural Language Math Proofs

Recent advances in large language models (LLMs) for mathematical reasoning have largely focused on tasks with easily verifiable final answers while generating and verifying natural language math proofs remains an open challenge. We identify the absence of a reliable, fine-grained evaluator for LLM-generated math proofs as a critical gap. To address this, we propose a systematic methodology for developing and validating evaluators that assign fine-grained scores on a 0-7 scale to model-generated math proofs. To enable this study, we introduce ProofBench, the first expert-annotated dataset of fine-grained proof ratings, spanning 145 problems from six major math competitions (USAMO, IMO, Putnam, etc) and 435 LLM-generated solutions from Gemini-2.5-Pro, o3, and DeepSeek-R1. Using ProofBench as a testbed, we systematically explore the evaluator design space across key axes: the backbone model, input context, instructions and evaluation workflow. Our analysis delivers ProofGrader, an evaluator that combines a strong reasoning backbone LM, rich context from reference solutions and marking schemes, and a simple ensembling method; it achieves a low Mean Absolute Error (MAE) of 0.926 against expert scores, significantly outperforming naive baselines. Finally, we demonstrate its practical utility in a best-of-$n$ selection task: at $n=16$, ProofGrader achieves an average score of 4.14/7, closing 78\% of the gap between a naive binary evaluator (2.48) and the human oracle (4.62), highlighting its potential to advance downstream proof generation.

cs.CL↗

arXiv2026

Delta Fair Sharing: Performance Isolation for Multi-Tenant Storage Systems

Modern storage systems, often deployed to support multiple tenants in the cloud, must provide performance isolation. Unfortunately, traditional approaches such as fair sharing do not provide performance isolation for storage systems, because their resources (e.g., write buffers and read caches) exhibit high preemption delays. These delays lead to unacceptable spikes in client tail latencies, as clients may be forced to wait arbitrarily long to receive their fair share of resources. We introduce Delta Fair Sharing, a family of algorithms for sharing resources with high preemption delays. These algorithms satisfy two key properties: $δ$-fairness, which bounds a client's delay in receiving its fair share of resources to $δ$ time units, and $δ$-Pareto-efficiency, which allocates unused resources to clients with unmet demand. Together, these properties capture resource-acquisition delays end-to-end, bound well-behaved clients' tail-latency spikes to $δ$ time units, and ensure high utilization. We implement such algorithms in FAIRDB, an extension of RocksDB. Our evaluation shows that FAIRDB isolates well-behaved clients from high-demand workloads better than state-of-the-art alternatives.

cs.DB↗

arXiv2025

Rate-Optimal Streaming Codes Under an Extended Delay Profile for Three-Node Relay Networks With Burst Erasures

This paper investigates streaming codes for three-node relay networks under burst packet erasures with a delay constraint $T$. In any sliding window of $T+1$ consecutive packets, the source-to-relay and relay-to-destination channels may introduce burst erasures of lengths at most $b_1$ and $b_2$, respectively. Let $u = \max\{b_1, b_2\}$ and $v = \min\{b_1, b_2\}$. Singhvi et al. proposed a construction achieving the optimal rate when $u\mid (T-u-v)$. In this paper, we present an extended delay profile method that attains the optimal rate under a relaxed constraint $\frac{T - u - v}{2u - v} \leq \left\lfloor \frac{T - u - v}{u} \right\rfloor$ and it strictly cover restriction $u\mid (T-u-v)$. %Furthermore, we demonstrate that the optimal rate for streaming codes is not achievable when $0< T-u-v<v$ under the convolutional code framework.

cs.IT↗

arXiv2025

Rate-Optimal Streaming Codes over Three-Node Relay Networks with Burst Erasures

This paper investigates streaming codes over three-node relay networks under burst packet erasures with a delay constraint $T$. In any sliding window of $T+1$ consecutive packets, the source-to-relay and relay-to-destination channels may introduce burst erasures of lengths at most $b_1$ and $b_2$, respectively. Singhvi et al. proposed a construction achieving the optimal code rate when $\max\{b_1,b_2\}\mid (T-b_1-b_2)$. We construct streaming codes with the optimal rate under the condition $T\geq b_1+b_2+\frac{b_1b_2}{|b_1-b_2|}$, thereby enriching the family of rate-optimal streaming codes for three-node relay networks.

cs.IT↗

arXiv2025

Review of Feed-forward 3D Reconstruction: From DUSt3R to VGGT

3D reconstruction, which aims to recover the dense three-dimensional structure of a scene, is a cornerstone technology for numerous applications, including augmented/virtual reality, autonomous driving, and robotics. While traditional pipelines like Structure from Motion (SfM) and Multi-View Stereo (MVS) achieve high precision through iterative optimization, they are limited by complex workflows, high computational cost, and poor robustness in challenging scenarios like texture-less regions. Recently, deep learning has catalyzed a paradigm shift in 3D reconstruction. A new family of models, exemplified by DUSt3R, has pioneered a feed-forward approach. These models employ a unified deep network to jointly infer camera poses and dense geometry directly from an Unconstrained set of images in a single forward pass. This survey provides a systematic review of this emerging domain. We begin by dissecting the technical framework of these feed-forward models, including their Transformer-based correspondence modeling, joint pose and geometry regression mechanisms, and strategies for scaling from two-view to multi-view scenarios. To highlight the disruptive nature of this new paradigm, we contrast it with both traditional pipelines and earlier learning-based methods like MVSNet. Furthermore, we provide an overview of relevant datasets and evaluation metrics. Finally, we discuss the technology's broad application prospects and identify key future challenges and opportunities, such as model accuracy and scalability, and handling dynamic scenes.

cs.CV↗

arXiv2025

Reasoning Models Can Be Effective Without Thinking

Recent LLMs have significantly improved reasoning capabilities, primarily by including an explicit, lengthy Thinking process as part of generation. In this paper, we question whether this explicit thinking is necessary. Using the state-of-the-art DeepSeek-R1-Distill-Qwen, we find that bypassing the thinking process via simple prompting, denoted as NoThinking, can be surprisingly effective. When controlling for the number of tokens, NoThinking outperforms Thinking across a diverse set of seven challenging reasoning datasets--including mathematical problem solving, formal theorem proving, and coding--especially in low-budget settings, e.g., 51.3 vs. 28.9 on ACM 23 with 700 tokens. Notably, the performance of NoThinking becomes more competitive with pass@k as k increases. Building on this observation, we demonstrate that a parallel scaling approach that uses NoThinking to generate N outputs independently and aggregates them is highly effective. For aggregation, we use task-specific verifiers when available, or we apply simple best-of-N strategies such as confidence-based selection. Our method outperforms a range of baselines with similar latency using Thinking, and is comparable to Thinking with significantly longer latency (up to 9x). Together, our research encourages a reconsideration of the necessity of lengthy thinking processes, while also establishing a competitive reference for achieving strong reasoning performance in low-budget settings or at low latency using parallel scaling.

cs.AI↗

arXiv2025

The Danger of Overthinking: Examining the Reasoning-Action Dilemma in Agentic Tasks

Large Reasoning Models (LRMs) represent a breakthrough in AI problem-solving capabilities, but their effectiveness in interactive environments can be limited. This paper introduces and analyzes overthinking in LRMs. A phenomenon where models favor extended internal reasoning chains over environmental interaction. Through experiments on software engineering tasks using SWE Bench Verified, we observe three recurring patterns: Analysis Paralysis, Rogue Actions, and Premature Disengagement. We propose a framework to study these behaviors, which correlates with human expert assessments, and analyze 4018 trajectories. We observe that higher overthinking scores correlate with decreased performance, with reasoning models exhibiting stronger tendencies toward overthinking compared to non-reasoning models. Our analysis reveals that simple efforts to mitigate overthinking in agentic environments, such as selecting the solution with the lower overthinking score, can improve model performance by almost 30% while reducing computational costs by 43%. These results suggest that mitigating overthinking has strong practical implications. We suggest that by leveraging native function-calling capabilities and selective reinforcement learning overthinking tendencies could be mitigated. We also open-source our evaluation framework and dataset to facilitate research in this direction at https://github.com/AlexCuadron/Overthinking.

cs.AI↗

arXiv2024

Enhancing Cyber-Resilience in Integrated Energy System Scheduling with Demand Response Using Deep Reinforcement Learning

Optimally scheduling multi-energy flow is an effective method to utilize renewable energy sources (RES) and improve the stability and economy of integrated energy systems (IES). However, the stable demand-supply of IES faces challenges from uncertainties that arise from RES and loads, as well as the increasing impact of cyber-attacks with advanced information and communication technologies adoption. To address these challenges, this paper proposes an innovative model-free resilience scheduling method based on state-adversarial deep reinforcement learning (DRL) for integrated demand response (IDR)-enabled IES. The proposed method designs an IDR program to explore the interaction ability of electricity-gas-heat flexible loads. Additionally, the state-adversarial Markov decision process (SA-MDP) model characterizes the energy scheduling problem of IES under cyber-attack, incorporating cyber-attacks as adversaries directly into the scheduling process. The state-adversarial soft actor-critic (SA-SAC) algorithm is proposed to mitigate the impact of cyber-attacks on the scheduling strategy, integrating adversarial training into the learning process to against cyber-attacks. Simulation results demonstrate that our method is capable of adequately addressing the uncertainties resulting from RES and loads, mitigating the impact of cyber-attacks on the scheduling strategy, and ensuring a stable demand supply for various energy sources. Moreover, the proposed method demonstrates resilience against cyber-attacks. Compared to the original soft actor-critic (SAC) algorithm, it achieves a 10% improvement in economic performance under cyber-attack scenarios.

eess.SY↗

arXiv2023

Deep Reinforcement Learning-driven Cross-Community Energy Interaction Optimal Scheduling

In order to coordinate energy interactions among various communities and energy conversions among multi-energy subsystems within the multi-community integrated energy system under uncertain conditions, and achieve overall optimization and scheduling of the comprehensive energy system, this paper proposes a comprehensive scheduling model that utilizes a multi-agent deep reinforcement learning algorithm to learn load characteristics of different communities and make decisions based on this knowledge. In this model, the scheduling problem of the integrated energy system is transformed into a Markov decision process and solved using a data-driven deep reinforcement learning algorithm, which avoids the need for modeling complex energy coupling relationships between multi-communities and multi-energy subsystems. The simulation results show that the proposed method effectively captures the load characteristics of different communities and utilizes their complementary features to coordinate reasonable energy interactions among them. This leads to a reduction in wind curtailment rate from 16.3% to 0% and lowers the overall operating cost by 5445.6 Yuan, demonstrating significant economic and environmental benefits.

eess.SY↗

arXiv2023

Context Sensitivity without Contexts: A Cut-Shortcut Approach to Fast and Precise Pointer Analysis

Over the past decades, context sensitivity has been considered as one of the most effective ideas for improving the precision of pointer analysis for Java. However, despite great precision benefits, as each method is equivalently cloned and analyzed under each context, context sensitivity brings heavy efficiency costs. In this work, we present a fundamentally different approach called Cut-Shortcut for fast and precise pointer analysis for Java. Its insight is simple: the main effect of cloning methods under different contexts is to filter spurious object flows that have been merged inside a callee method; from the view of a typical pointer flow graph (PFG), such effect can be simulated by cutting off (Cut) the edges that introduce precision loss to certain pointers and adding Shortcut edges directly from source pointers to the target ones circumventing the method on PFG. As a result, we can achieve the effect of context sensitivity without contexts. We identify three general program patterns and develop algorithms based on them to safely cut off and add shortcut edges on PFG, formalize them and formally prove the soundness. To comprehensively validate Cut-Shortcut's effectiveness, we implement two versions of Cut-Shortcut for two state-of-the-art pointer analysis frameworks for Java, one in Datalog for the declarative Doop and the other in Java for the imperative Tai-e, and we consider all the large and complex programs used in recent literatures that meet the experimental requirements. The evaluation results are extremely promising: Cut-Shortcut is even able to run faster than context insensitivity for most evaluated programs while obtaining high precision that is comparable to context sensitivity (if scalable) in both frameworks. This is for the first time that we have been able to achieve such a good efficiency and precision trade-off for those hard-to-analyze programs.

cs.PL↗