Searcharxiv⌕ Search

arXiv subjects

Zhi Jin

Publications and source records attributed to Zhi Jin.

At least 37 records · Page 2Linked to original sources
arXiv2026

On the Road to Personalized Code Intelligence: Portraiting and Assisting Developers Based on Their In-IDE Behaviors

With the advent of large language models, research in automated software engineering has increasingly focused on leveraging these models to achieve a deeper semantic understanding of code or to engineer sophisticated agent-based processes. However, this research trajectory has largely overlooked a critical factor: the developers themselves. Programming is a deeply individualized activity; developers exhibit significant variation in their tool-chain preferences, domain-specific expertise, and problem-solving strategies. Consequently, the current paradigm of one-size-fits-all code intelligence systems struggles to accommodate the needs of individual developers. To address this gap, we introduce VirtualME, a novel IDE-embedded data infrastructure designed to model the developer by continuously capturing and interpreting their dynamic programming behaviors and preferences. VirtualME contains three components. (1) Log-level Behavior Extraction: it captures and extracts developers' log-level behaviors from IDE. (2) Task-level Behavior Recognition: it aggregates log-level behaviors into task-level behaviors via a multi-agent pipeline. (3) Developer-personality Measurement: it builds a rule engine to distill a four-dimensional developer persona: technology stack, ability, behavioral habits, and learning style. On top of VirtualME, we propose a solution for personalized repository-level knowledge Q&A by integrating the developer persona into the Q&A agent. We evaluated VirtualME by building a multi-repository benchmark with real-world developer trajectories, balancing correctness and personalization. Experimental results show that VirtualME-enhanced answers outperform generic baselines on five dimensions, yielding an average 33.80% improvement. Our results demonstrate that abundant, continuous developer-behavior data can pave the new way for adaptive and personalized code intelligence.

cs.SE↗
arXiv2026

MARR: Module-Adaptive Residual Reconstruction for Low-Bit Post-Training Quantization

Recently, residual reconstruction-based model quantization methods have achieved promising performance in low-bit post-training quantization (PTQ) by introducing cross-layer residuals to reduce error accumulated from previous layers.However, these residuals may also introduce additional bias arising from the Hessian-approximation (HA) assumption underlying reconstruction-based PTQ, leading to suboptimal quantization performance.In this work, we analyze that multiplying the residual term by a scaling coefficient provides a direct way to mitigate the HA bias associated with residual strength, while preserving accumulated-error correction. More importantly, we observe that this trade-off is module-dependent, making a single global residual strength insufficient to balance effective correction and residual-related bias across modules.Based on these observations, we propose Module-Adaptive Residual Reconstruction (MARR), which assigns a module-specific scaling coefficient to adaptively balance accumulated-error correction and residual-related HA bias for each module.To avoid expensive per-module coefficient search and obtain a stable coefficient estimate, we design a Proportional-Integral-Derivative (PID)-based adaptive update strategy that uses reconstruction error as feedback to progressively refine this coefficient. Experiments on several typical large language models (LLMs) and vision transformers (ViTs) demonstrate the effectiveness of MARR under low-bit quantization (less than or equal to 4-bit), achieving up to 20.2% performance gains on LLMs and up to 4.6% relative gains on ViTs over the residual reconstruction state-of-the-art methods.Code will be made publicly available upon acceptance.

cs.LG↗
arXiv2026

Low Light Image Enhancement Challenge at NTIRE 2026

This paper presents a comprehensive review of the NTIRE 2026 Low Light Image Enhancement Challenge, highlighting the proposed solutions and final results. The objective of this challenge is to identify effective networks capable of producing clearer and visually compelling images in diverse and challenging conditions by learning representative visual cues with the purpose of restoring information loss due to low-contrast and noisy images. A total of 195 participants registered for the first track and 153 for the second track of the competition, and 22 teams ultimately submitted valid entries. This paper thoroughly evaluates the state-of-the-art advances in (joint denoising and) low-light image enhancement, showcasing the significant progress in the field, while leveraging samples of our novel dataset.

cs.CV↗
arXiv2026

From I/O to Code with Discovery Agent

The automatic synthesis of a program from any form of specification is regarded as a holy grail of computer science. Fueled by LLMs, NL2Code has achieved tremendous success, yet the fundamentally more challenging task of synthesizing programs from input-output behavior, which we refer to as IO2Code, remains largely unsolved. Whereas NL2Code can exploit the semantic alignment between natural language and code acquired during pretraining, IO2Code requires recovering underlying principles from concrete computational behavior, navigating a vast and underspecified hypothesis space. To address this, we propose DIO-Agent, a discovery agent for IO2Code. Our method frames IO2Code as an evolutionary search over discrete program space, in which an LLM serves as the mutation operator and concrete error signals from execution guide each mutation. To prevent the search from wandering into structurally complex yet incorrect dead ends, we introduce the Transformation Priority Premise as a mutation prior that biases the LLM toward the simplest hypothesis consistent with current evidence, progressively escalating from constants to conditionals to iteration only when simpler constructs are insufficient. To facilitate systematic study, we further construct an IO2CodeBench spanning multiple difficulty levels. Extensive experiments show that DIO-Agent consistently outperforms both traditional program-by-example method and SOTA evolution-agent baselines across all difficulty levels and various LLMs, while substantially surpassing test-time scaling strategies with equivalent sampling budgets.

cs.LG↗
arXiv2026

From Chat to Interview: Agentic Requirements Elicitation with an Experience Ontology

Requirements elicitation interviews are crucial and time-consuming in requirements engineering, but heavily rely on the experience of requirements analysts. Although recent advancements in large language models (LLMs) have created new opportunities to automate this process, existing approaches rely solely on LLMs for free-form chat without taking into account the interview and development experience. That leads to the omission of implicit requirements and redundant questions. Practically, experienced analysts implicitly follow a structured cognitive framework when conducting requirements elicitation. Inspired by this observation, this paper proposes an interview agent named OntoAgent for the elicitation of requirements guided by an experience ontology. OntoAgent automatically analyzes domain-specific requirements descriptions to construct an experience ontology, which organizes requirements concerns into an ontology to support systematic and explainable interviews. During the interview, OntoAgent first performs four operations (i.e., ParseUser, ScoreOnto, ReRankOnto, GatePrune) guided by the ontology to identify the relevant requirement concerns. The selected concern is then combined with the current dialogue context to generate the elicitation question. To validate OntoAgent, we conduct comprehensive quantitative experiments using the widely adopted website application domain. The results show that OntoAgent significantly outperforms existing baselines in both elicitation effectiveness and questioning efficiency, achieving a 33% improvement in IRE and a 21% improvement in TKQR. Ablation studies further validate the contribution of each key design component. In addition, a qualitative user study demonstrates its practical advantages in real-world scenarios. We believe that OntoAgent can also be extended to requirements interview tasks in other domains.

cs.SE↗
arXiv2026

ClarifySTL: An Interactive LLM Agent Framework for STL Transformation through Requirements Clarification

Signal Temporal Logic (STL) is a formal language for specifying real-time behaviors of cyber-physical systems (CPS). Automatically transforming natural language requirements into STL specifications has received growing attention. Recent efforts leveraging large language models (LLMs) have demonstrated impressive performance, but some natural language requirements in practice contain vague or ambiguous information, which remains challenging for LLMs to handle. To address these challenges, we propose ClarifySTL, an interactive LLM-agent framework that enhances STL transformation through requirements clarification. ClarifySTL first detects vague expressions that indicate underspecified information in a requirement. If any vagueness is detected, it generates targeted clarification queries to guide users in supplementing the requirement until all necessary details are provided. Subsequently, if ClarifySTL detects ambiguities, it formulates focused ambiguity clarification queries and updates the requirements based on user feedback until all ambiguities are resolved. Finally, the requirements with vagueness and ambiguity clarified are transformed into STL specifications using LLMs. This interactive framework ensures that the resulting STL formulas faithfully capture user intent while reducing the burden on the user. We evaluate ClarifySTL on the representative benchmarks DeepSTL and STL-DivEn, as well as our newly introduced AmbiEval benchmark, which is specifically designed to assess the performance of the agents in handling vagueness and ambiguity, including both detection and query generation. The experimental results show that ClarifySTL is effective.

cs.SE↗
arXiv2026

Saber: An Efficient Sampling with Adaptive Acceleration and Backtracking Enhanced Remasking for Diffusion Language Model

Diffusion language models (DLMs) are emerging as a compelling alternative to the dominant autoregressive paradigm, offering inherent advantages in parallel generation and bidirectional context modeling. However, for the tasks with strict structural constraints such as code generation, DLMs face a critical trade-off between inference speed and output quality, where accelerating generation by reducing sampling steps often leads to catastrophic performance collapse. We find that the fundamental reasons are: 1) the generation difficulty is non-uniform in the structured sequence decoding steps, making DLM's static acceleration strategy suboptimal; 2) the context of tokens generated by DLM evolves continuously, causing early high-confidence predictions to turn into irreversible errors. In this paper, we introduce efficient Sampling with Adaptive acceleration and Backtracking Enhanced Remasking (i.e., Saber), a novel training-free sampling algorithm for DLMs that first achieves both better inference speed and output quality in code generation. Saber dynamically adjusts the number of tokens unmasked per step based on the model's evolving confidence, and utilizes a backtracking mechanism to revert tokens whose confidence drops as new context emerges, with its effectiveness supported by theoretical analysis. Extensive experiments on multiple mainstream code generation benchmarks show that Saber boosts Pass@1 accuracy by an average of 1.9\% over mainstream DLM sampling methods, while achieving an average 251.4\% inference speedup. By leveraging the inherent advantages of DLMs, our work significantly narrows the performance gap with autoregressive models in code generation.

cs.AI↗
arXiv2026

Think Anywhere in Code Generation

Recent advances in reasoning Large Language Models (LLMs) have primarily relied on upfront thinking, where reasoning occurs before final answer. However, this approach suffers from critical limitations in code generation, where upfront thinking is often insufficient as problems' full complexity only reveals itself during code implementation. Moreover, it cannot adaptively allocate reasoning effort throughout the code generation process where difficulty varies significantly. In this paper, we propose Think-Anywhere, a novel reasoning mechanism that enables LLMs to invoke thinking on-demand at any token position during code generation. We achieve Think-Anywhere by first teaching LLMs to imitate the reasoning patterns through cold-start training, then leveraging outcome-based RL rewards to drive the model's autonomous exploration of when and where to invoke reasoning. Extensive experiments on four mainstream code generation benchmarks (i.e., LeetCode, LiveCodeBench, HumanEval, and MBPP) show that Think-Anywhere achieves state-of-the-art performance over both existing reasoning methods and recent post-training approaches, while demonstrating consistent generalization across diverse LLMs. Our analysis further reveals that Think-Anywhere enables the model to adaptively invoke reasoning at high-entropy positions, providing enhanced interpretability.

cs.SE↗
arXiv2026

KOCO-BENCH: Can Large Language Models Leverage Domain Knowledge in Software Development?

Large language models (LLMs) excel at general programming but struggle with domain-specific software development, necessitating domain specialization methods for LLMs to learn and utilize domain knowledge and data. However, existing domain-specific code benchmarks cannot evaluate the effectiveness of domain specialization methods, which focus on assessing what knowledge LLMs possess rather than how they acquire and apply new knowledge, lacking explicit knowledge corpora for developing domain specialization methods. To this end, we present KOCO-BENCH, a novel benchmark designed for evaluating domain specialization methods in real-world software development. KOCO-BENCH contains 6 emerging domains with 11 software frameworks and 25 projects, featuring curated knowledge corpora alongside multi-granularity evaluation tasks including domain code generation (from function-level to project-level with rigorous test suites) and domain knowledge understanding (via multiple-choice Q&A). Unlike previous benchmarks that only provide test sets for direct evaluation, KOCO-BENCH requires acquiring and applying diverse domain knowledge (APIs, rules, constraints, etc.) from knowledge corpora to solve evaluation tasks. Our evaluations reveal that KOCO-BENCH poses significant challenges to state-of-the-art LLMs. Even with domain specialization methods (e.g., SFT, RAG, kNN-LM) applied, improvements remain marginal. Best-performing coding agent, Claude Code, achieves only 34.2%, highlighting the urgent need for more effective domain specialization methods. We release KOCO-BENCH, evaluation code, and baselines to advance further research at https://github.com/jiangxxxue/KOCO-bench.

cs.SE↗
arXiv2026

CodeRL+: Improving Code Generation via Reinforcement with Execution Semantics Alignment

While Large Language Models (LLMs) excel at code generation by learning from vast code corpora, a fundamental semantic gap remains between their training on textual patterns and the goal of functional correctness, which is governed by formal execution semantics. Reinforcement Learning with Verifiable Rewards (RLVR) approaches attempt to bridge this gap using outcome rewards from executing test cases. However, solely relying on binary pass/fail signals is inefficient for establishing a well-aligned connection between the textual representation of code and its execution semantics, especially for subtle logical errors within the code. In this paper, we propose CodeRL+, a novel approach that integrates execution semantics alignment into the RLVR training pipeline for code generation. CodeRL+ enables the model to infer variable-level execution trajectory, providing a direct learning signal of execution semantics. CodeRL+ can construct execution semantics alignment directly using existing on-policy rollouts and integrates seamlessly with various RL algorithms. Extensive experiments demonstrate that CodeRL+ outperforms post-training baselines (including RLVR and Distillation), achieving a 4.6% average relative improvement in pass@1. CodeRL+ generalizes effectively to other coding tasks, yielding 15.5% and 4.4% higher accuracy on code-reasoning and test-output-generation benchmarks, respectively. CodeRL+ shows strong applicability across diverse RL algorithms and LLMs. Furthermore, probe analyses provide compelling evidence that CodeRL+ strengthens the alignment between code's textual representations and its underlying execution semantics.

cs.SE↗
arXiv2026

Bridging the Gap between User Intent and LLM: A Requirement Alignment Approach for Code Generation

Code generation refers to automatically producing executable programs from user requirements. Recently, researchers have explored approaches to enhance the correctness of generated code with advanced large language models. Although achieving improvements, existing approaches focus on designing reasoning strategies or post-refinement methods to enhance code generation performance. Despite their differences, all these methods share a common assumption: the LLM can correctly understand the given requirement. However, this assumption does not always hold. To fill this gap, we propose REA-Coder, a requirement alignment approach to enhance the code generation performance of LLMs. REA-Coder involves first identifying the requirement content that does not align with LLMs and aligning the requirements. Then, based on the aligned requirements, LLMs generate code and further verify whether the generated code aligns with the requirements, iterating this process of requirement alignment and code generation until generating correct code or achieving the maximum number of iterations. Experimental results show that REA-Coder outperforms all advanced baselines on four LLMs across five programming benchmarks. Concretely, REA-Coder achieves average improvements of 7.93%, 30.25%, 26.75%, 8.59%, and 8.64% on the five benchmark datasets, demonstrating the effectiveness of requirement alignment for improving the code generation performance of LLMs.

cs.SE↗
arXiv2026

RL-PLUS: Countering Capability Boundary Collapse of LLMs in Reinforcement Learning with Hybrid-policy Optimization

Reinforcement Learning with Verifiable Reward (RLVR) has significantly advanced the complex reasoning abilities of Large Language Models (LLMs). However, it struggles to break through the inherent capability boundaries of the base LLM, due to its essentially on-policy strategy coupled with LLM's immense action space and sparse reward. Critically, RLVR can lead to the capability boundary collapse, narrowing the LLM's problem-solving scope. To address this problem, we propose RL-PLUS, a novel hybrid-policy optimization approach for LLMs that synergizes internal exploitation with external data to achieve stronger reasoning capabilities and surpass the boundaries of base models. RL-PLUS integrates two core components, i.e., Multiple Importance Sampling to address distributional mismatch from external data, and Exploration-Based Advantage Function to guide the model towards high-value, unexplored reasoning paths. We provide both theoretical analysis and extensive experiments to demonstrate the superiority and generalizability of our approach. Compared with existing RLVR methods, RL-PLUS achieves 1) state-of-the-art performance on six math reasoning benchmarks; 2) superior performance on six out-of-distribution reasoning tasks; 3) consistent and significant gains across diverse model families, with average relative improvements up to 69.2\%. Moreover, the analysis of Pass@k curves indicates that RL-PLUS effectively resolves the capability boundary collapse problem.

cs.AI↗
arXiv2026

Evaluating the Formal Reasoning Capabilities of Large Language Models through Chomsky Hierarchy

The formal reasoning capabilities of LLMs are crucial for advancing automated software engineering. However, existing benchmarks for LLMs lack systematic evaluation based on computation and complexity, leaving a critical gap in understanding their formal reasoning capabilities. Therefore, it is still unknown whether SOTA LLMs can grasp the structured, hierarchical complexity of formal languages as defined by Computation Theory. To address this, we introduce ChomskyBench, a benchmark for systematically evaluating LLMs through the lens of Chomsky Hierarchy. Unlike prior work that uses vectorized classification for neural networks, ChomskyBench is the first to combine full Chomsky Hierarchy coverage, process-trace evaluation via natural language, and deterministic symbolic verifiability. ChomskyBench is composed of a comprehensive suite of language recognition and generation tasks designed to test capabilities at each level. Extensive experiments indicate a clear performance stratification that correlates with the hierarchy's levels of complexity. Our analysis reveals a direct relationship where increasing task difficulty substantially impacts both inference length and performance. Furthermore, we find that while larger models and advanced inference methods offer notable relative gains, they face severe efficiency barriers: achieving practical reliability would require prohibitive computational costs, revealing that current limitations stem from inefficiency rather than absolute capability bounds. A time complexity analysis further indicates that LLMs are significantly less efficient than traditional algorithmic programs for these formal tasks. These results delineate the practical limits of current LLMs, highlight the indispensability of traditional software tools, and provide insights to guide the development of future LLMs with more powerful formal reasoning capabilities.

cs.CL↗
arXiv2026

Unifying Heterogeneous Degradations: Uncertainty-Aware Diffusion Bridge Model for All-in-One Image Restoration

All-in-One Image Restoration (AiOIR) faces the fundamental challenge in reconciling conflicting optimization objectives across heterogeneous degradations. Existing methods are often constrained by coarse-grained control mechanisms or fixed mapping schedules, yielding suboptimal adaptation. To address this, we propose an Uncertainty-Aware Diffusion Bridge Model (UDBM), which innovatively reformulates AiOIR as a stochastic transport problem steered by pixel-wise uncertainty. By introducing a relaxed diffusion bridge formulation which replaces the strict terminal constraint with a relaxed constraint, we model the uncertainty of degradations while theoretically resolving the drift singularity inherent in standard diffusion bridges. Furthermore, we devise a dual modulation strategy: the noise schedule aligns diverse degradations into a shared high-entropy latent space, while the path schedule adaptively regulates the transport trajectory motivated by the viscous dynamics of entropy regularization. By effectively rectifying the transport geometry and dynamics, UDBM achieves state-of-the-art performance across diverse restoration tasks within a single inference step.

cs.CV↗
arXiv2026

Your Inference Request Will Become a Black Box: Confidential Inference for Cloud-based Large Language Models

The increasing reliance on cloud-hosted Large Language Models (LLMs) exposes sensitive client data, such as prompts and responses, to potential privacy breaches by service providers. Existing approaches fail to ensure privacy, maintain model performance, and preserve computational efficiency simultaneously. To address this challenge, we propose Talaria, a confidential inference framework that partitions the LLM pipeline to protect client data without compromising the cloud's model intellectual property or inference quality. Talaria executes sensitive, weight-independent operations within a client-controlled Confidential Virtual Machine (CVM) while offloading weight-dependent computations to the cloud GPUs. The interaction between these environments is secured by our Reversible Masked Outsourcing (ReMO) protocol, which uses a hybrid masking technique to reversibly obscure intermediate data before outsourcing computations. Extensive evaluations show that Talaria can defend against state-of-the-art token inference attacks, reducing token reconstruction accuracy from over 97.5% to an average of 1.34%, all while being a lossless mechanism that guarantees output identical to the original model without significantly decreasing efficiency and scalability. To the best of our knowledge, this is the first work that ensures clients' prompts and responses remain inaccessible to the cloud, while also preserving model privacy, performance, and efficiency.

cs.CR↗
arXiv2026

Toward a CMOS-integrated quantum diamond biosensor based on NV centers

We report progress toward a CMOS-integrated quantum diamond biosensing platform that combines nitrogen-vacancy (NV) centers in diamond with a custom 40 nm CMOS Single-Photon Avalanche Diode (SPAD) array. The system integrates on-chip active quenching and digital readout with external FPGA-based photon counting, compact microwave delivery, and practical optical excitation and collection schemes to support widefield optically detected magnetic resonance (ODMR). System-level design considerations spanning fluorescence collection efficiency, detector count-rate capability, and microwave homogeneity are analyzed with biological compatibility and scalability in mind. Using superparamagnetic iron oxide nanoparticle (SPION)-labeled HEK293T cells as a representative use case, simple dipole-field estimates indicate that sub-$μ$T sensitivity is required to resolve ODMR shifts within typical ensemble linewidths. Based on the proposed architecture and efficiency analysis, a magnetic field sensitivity of approximately 90 nT/$\sqrt{\mathrm{Hz}}$ per pixel is estimated. These results outline a practical path from optics-heavy quantum diamond microscopes toward compact, CMOS-integrated NV-based biosensors for quantitative magnetic imaging in complex biological environments.

physics.app-ph↗
arXiv2026

ReqElicitGym: An Evaluation Environment for Interview Competence in Conversational Requirements Elicitation

With the rapid improvement of LLMs' coding capabilities, the bottleneck of LLM-based automated software development is shifting from generating correct code to eliciting users' requirements. Despite growing interest, the interview competence of LLMs in conversational requirements elicitation remains fully underexplored. Existing evaluations often depend on a few scenarios, real user interaction, and subjective human scoring, which hinders systematic and quantitative comparison. To address these challenges, we propose ReqElicitGym, an interactive and automatic evaluation environment for assessing interview competence in conversational requirements elicitation. Specifically, ReqElicitGym introduces a new evaluation dataset and designs both an interactive oracle user and a task evaluator. The dataset contains 101 website requirements elicitation scenarios spanning 10 application types. Both the oracle user and the task evaluator achieve high agreement with real users and expert judgment. Using our ReqElicitGym, any automated conversational requirements elicitation approach (e.g., LLM-based agents) can be evaluated in a reproducible and quantitative manner through interaction with the environment. Based on our ReqElicitGym, we conduct a systematic empirical study on seven representative LLMs, and the results show that current LLMs still exhibit limited interview competence in uncovering implicit requirements. Particularly, they elicit less than half of the users' implicit requirements, and their effective elicitation questions often emerge in later turns of the dialogue. Besides, we found LLMs can elicit interaction and content implicit requirements, but consistently struggle with style-related requirements. We believe ReqElicitGym will facilitate the evaluation and development of automated conversational requirements elicitation.

cs.SE↗
arXiv2026

aiXcoder-7B-v2: Training LLMs to Fully Utilize the Long Context in Repository-level Code Completion

Large Language Models (LLMs) have shown promising results in repository-level code completion, which completes code based on the in-file and cross-file context of a repository. The cross-file context typically contains different types of information (e.g., relevant APIs and similar code) and is lengthy. In this paper, we found that LLMs struggle to fully utilize the information in the cross-file context. We hypothesize that one of the root causes of the limitation is the misalignment between pre-training (i.e., relying on nearby context) and repo-level code completion (i.e., frequently attending to long-range cross-file context). To address the above misalignment, we propose Code Long-context Alignment - COLA, a purely data-driven approach to explicitly teach LLMs to focus on the cross-file context. Specifically, COLA constructs a large-scale repo-level code completion dataset - COLA-132K, where each sample contains the long cross-file context (up to 128K tokens) and requires generating context-aware code (i.e., cross-file API invocations and code spans similar to cross-file context). Through a two-stage training pipeline upon COLA-132K, LLMs learn the capability of finding relevant information in the cross-file context, thus aligning LLMs with repo-level code completion. We apply COLA to multiple popular LLMs (e.g., aiXcoder-7B) and extensive experiments on COLA-132K and a public benchmark - CrossCodeEval. Our experiments yield the following results. 1) Effectiveness. COLA substantially improves the performance of multiple LLMs in repo-level code completion. For example, it improves aiXcoder-7B by up to 19.7% in exact match. 2) Generalizability. The capability learned by COLA can generalize to new languages. 3) Enhanced Context Utilization Capability. We design two probing experiments, which show COLA improves the capability of LLMs in utilizing the information in cross-file context.

cs.SE↗