Papers.
-
2026 Under reviewFormalizing Latent Thoughts: Four Axioms of Thought Representation in LLMs
We formalize four axioms for LLM thought representations and show structural failures in candidate representations.
-
2026 Under reviewARISE: A Repository-level Graph Representation and Toolset for Agentic Program Repair
Builds a multi-granularity repository graph with statement-level data-flow slicing so a program-repair agent can trace where a variable is defined and used in a single call. Mounted on SWE-agent, it resolves 4.7 more points of SWE-bench Lite issues and nearly doubles line-level fault-localization recall.
-
2026 TOSEM ACM Transactions on Software Engineering and Methodology Under reviewContext-Augmented Code Generation Using Programming Knowledge Graphs
Grounds code generation in a programming knowledge graph so the model retrieves precise structural context (call sites, types, idioms) before generating, instead of relying on lossy in-context recall.
-
2026 FSE ACM International Conference on the Foundations of Software EngineeringPanther: Faster and Cheaper Computations with Randomized Numerical Linear Algebra
Randomized numerical linear algebra primitives that drop into existing PyTorch layers – 5x speedup over the dense baseline and 75% parameter reduction at negligible accuracy cost on BERT.
-
2025 EMSE Empirical Software Engineering Under reviewAnalysis of AdvFusion: Adapter-based Multilingual Learning for Code LLMs
An empirical analysis of AdvFusion, an adapter-based multilingual training scheme for code LLMs, with a focus on cross-language transfer and pareto trade-offs.
-
2023 ICECCE International Conference on Electrical, Communication and Computer EngineeringAGS: Arabic GPT Summarization Corpus
An Arabic abstractive-summarisation dataset built via prompt engineering – the first Arabic abstractive corpus labelled by an LLM. Backed our 1st-place AIC-1 (ICMTC) submission.
3D model: "3D Origami crane" by JuanG3D · CC BY 4.0