arXiv subjects

Tingyu Guo

Publications and source records attributed to Tingyu Guo.

3 recordsLinked to original sources

arXiv2026

ActionSplice: In-Flight Action Editing for Interactive World Models

Chunk-autoregressive video world models typically condition each generated chunk on one action. An action received during sampling must therefore wait for the next chunk, condition future solver evaluations on a state produced under the previous action, or trigger rollback that repeats completed evaluations. We introduce ActionSplice, an inference framework that formulates this problem as Counterfactual State Transport (CST). A lightweight corrector transports the interrupted backbone-native representation toward the matched state induced by the revised action at the same solver step. The world model and sampler remain frozen, and sampling resumes without replaying completed evaluations. The retargeting variant $\mathrm{CST}*{R}$ updates the entire active chunk, while the temporal-splicing variant $\mathrm{CST}*{T}$ preserves a temporal prefix and updates only the suffix. Across minWM-Wan Action2V and HY-WM1.5, $\mathrm{CST}*{R}$ reduces rollback-relative LPIPS by 61.5% and 75.9% relative to direct condition swapping. $\mathrm{CST}*{T}$ reduces suffix LPIPS by 56.1% and 77.5%, respectively, while providing $2.73\times$ and $1.69\times$ pixel-ready speedups over waiting. Under the HY-WorldPlay protocol, $\mathrm{CST}_{R}$ obtains a PSNR of 25.66 dB, an SSIM of 0.6902, and an LPIPS of 0.1337 against the original rollout.

cs.CV↗

arXiv2025

CoCoEvo: Co-Evolution of Programs and Test Cases to Enhance Code Generation

Large Language Models (LLMs) have shown remarkable performance in automated code generation. However, existing approaches often rely heavily on pre-defined test cases, which become impractical in scenarios where such cases are unavailable. While prior works explore filtering techniques between programs and test cases, they overlook the refinement of test cases. To address this limitation, we introduce CoCoEvo, a novel LLM-based co-evolution framework that simultaneously evolves programs and test cases. CoCoEvo eliminates the dependency on pre-defined test cases by generating both programs and test cases directly from natural language problem descriptions and function headers. The framework employs specialized evolutionary operators, including LLM-based crossover and mutation operators for program evolution, along with an additional test case generation operator for test case evolution. Additionally, we propose optimization strategies such as a crossover rate scheduler to balance exploration and convergence, and a multi-objective optimization method for test case selection. Experimental results on multiple state-of-the-art LLMs demonstrate that CoCoEvo surpasses existing methods, achieving state-of-the-art performance in automated code generation and testing. These results underscore the potential of co-evolutionary techniques in advancing the field of automated programming.

cs.SE↗

arXiv2024

Revisiting Evolutionary Program Repair via Code Language Model

Software defects are an inherent part of software development and maintenance. To address these defects, Automated Program Repair (APR) has been developed to fix bugs automatically. With the advent of Large Language Models, Code Language Models (CLMs) trained on code corpora excels in code generation, making them suitable for APR applications. Despite this progress, a significant limitation remains: many bugs necessitate multi-point edits for repair, yet current CLM-based APRs are restricted to single-point bug fixes, which severely narrows the scope of repairable bugs. Moreover, these tools typically only consider the direct context of the buggy line when building prompts for the CLM, leading to suboptimal repair outcomes due to the limited information provided. This paper introduces a novel approach, ARJA-CLM, which integrates the multiobjective evolutionary algorithm with CLM to fix multilocation bugs in Java projects. We also propose a context-aware prompt construction stratege, which enriches the prompt with additional information about accessible fields and methods for the CLM generating candidate statements. Our experiments on the Defects4J and APR-2024 competition benchmark demonstrate that ARJA-CLM surpasses many state-of-the-art repair systems, and performs well on multi-point bugs. The results also reveal that CLMs effectively utilize the provided field and method information within context-aware prompts to produce candidate statements.

cs.SE↗