arXiv subjects

Zijian Luo

Publications and source records attributed to Zijian Luo.

3 recordsLinked to original sources

arXiv2026

Update from Hell: Can Coding Agents Survive Hidden Breakage in Dependency Upgrades?

Modern software systems rely heavily on third-party dependencies, but upgrading those dependencies remains a costly maintenance activity. Dependency upgrades do not always preserve the function signatures, type systems, APIs, or runtime semantics assumed by existing code. Consequently, developers often need to perform source code adaptations to accommodate dependency-induced changes. However, such code-level changes are often not explicitly communicated to project maintainers, posing a significant challenge to software reliability. Meanwhile, coding agents have emerged as a new form of software development tool and are increasingly adopted by developers due to their automation capabilities. In this paper, we introduce DEPBENCH, a benchmark consisting of 203 real-world dependency-upgrade tasks across five package ecosystems spanning five language communities, each involving hidden code-level changes that require source code adaptation. We evaluate mainstream coding agents on DEPBENCH. The best completed configuration solves only 104/203 tasks (51.2%), with substantial variation across agent harnesses, models, and ecosystems, highlighting an important gap between current agent capabilities and real-world software maintenance needs.

cs.SE↗

arXiv2025

HiGR: Industrial-Scale Hierarchical Generative Slate Recommendation Framework in Tencent

Slate recommendation, which presents users with a ranked item list in a single display, is ubiquitous across mainstream online platforms. While recent generative recommendation methods have shown strong potential in modeling item sequences with semantic IDs, directly applying them to industrial-scale slate recommendation faces a fundamental disconnect: entangled SID spaces confound high-level list planning, fine-grained autoregressive decoding over long sequences limits semantic planning efficiency, and token-level objectives misalign with holistic slate quality. In this paper, we propose HiGR, an industrial-scale hierarchical generative framework for slate recommendation that bridges this disconnect through a co-designed pipeline. First, HiGR learns structured SIDs via a Prefix-Contrastive Residual Quantized VAE (PCRQ-VAE). By enforcing high-level prefixes to capture shared semantics, PCRQ-VAE creates a controllable discrete space that acts as a prerequisite for efficient planning. Leveraging this structured space, our Hierarchical Slate Decoder (HSD) shifts autoregressive modeling from entangled token-level decoding to coarse-grained preference embeddings. This design significantly reduces inference latency while allowing explicit global slate structure planning. Finally, this stable planning space enables an ORPO-based listwise alignment mechanism to optimize triple-objective implicit feedback-ranking fidelity, genuine user interest, and diversity. Extensive offline experiments show that HiGR outperforms state-of-the-art baselines by over 10% in offline recommendation quality while achieving a $5\times$ inference speedup. Online A/B tests on Tencent platforms further improve watch time by 1.22% and video plays by 1.73%. HiGR has been deployed on multiple Tencent platform surfaces, serving hundreds of millions of users and proving its industrial-scale applicability.

cs.IR↗

arXiv2025

Assessing Reliability of Statistical Maximum Coverage Estimators in Fuzzing

Background: Fuzzers are often guided by coverage, making the estimation of maximum achievable coverage a key concern in fuzzing. However, achieving 100% coverage is infeasible for most real-world software systems, regardless of effort. While static reachability analysis can provide an upper bound, it is often highly inaccurate. Recently, statistical estimation methods based on species richness estimators from biostatistics have been proposed as a potential solution. Yet, the lack of reliable benchmarks with labeled ground truth has limited rigorous evaluation of their accuracy. Objective: This work examines the reliability of reachability estimators from two axes: addressing the lack of labeled ground truth and evaluating their reliability on real-world programs. Methods: (1) To address the challenge of labeled ground truth, we propose an evaluation framework that synthetically generates large programs with complex control flows, ensuring well-defined reachability and providing ground truth for evaluation. (2) To address the criticism from use of synthetic benchmarks, we adapt a reliability check for reachability estimators on real-world benchmarks without labeled ground truth -- by varying the size of sampling units, which, in theory, should not affect the estimate. Results: These two studies together will help answer the question of whether current reachability estimators are reliable, and defines a protocol to evaluate future improvements in reachability estimation.

cs.SE↗