ai4se.codelab.shFall 2026
MATERIALS

Week by week.

WEEK 1

LLMs for Code — Foundations & Evaluation

Tokenization, pretraining incl. fill-in-the-middle, the model landscape, pass@k, and contamination-aware evaluation.

WEEK 2

Autonomous SE Agents — & SWE-bench

The ReAct loop, agent anatomy and the Agent–Computer Interface, SWE-bench (Verified/Live/Pro), and reading results honestly.

WEEK 3

Automated Program Repair

Localize → generate → validate; GenProg / SemFix / Angelix; the overfitting problem; neural and LLM-based repair.

WEEK 4

Neural Code Representations

Tokens vs AST/graph; CodeBERT, GraphCodeBERT, CodeT5, UniXcoder; contrastive pretraining and code retrieval.

WEEK 5

LLM-based Test Generation

The oracle problem; search-based (EvoSuite) to LLM-based (TestPilot, CoverUp, TestGen-LLM); coverage vs mutation.

WEEK 6

Code Review, Summarization & Documentation

Code → words a developer trusts: automated review (CodeReviewer), summarization, documentation, and the metric problem (BLEU vs usefulness).

WEEK 7

Vulnerability Detection & Secure Code

Learning-based detection (Devign/LineVul), the lab-to-reality gap, secure code generation ("Asleep at the Keyboard"), and hybrid analysis + LLM.