Program (Tentative) — RigoLLM 2026

9:00–9:10

Opening Remarks

Cheng Tan (Northeastern University) and Youshan Miao (Microsoft Research Asia)

9:10–10:10

Keynote I

Haibo Chen — Title TBD

Session Chair: TBD

10:10–10:40

Coffee Break

10:40–11:45

Session 1: Determinism & Reproducibility

Session Chair: TBD

10:40–10:55
Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy High-Impact Work (MICRO)
Peichen Xie (Microsoft Research)
10:55–11:10
Vitamin-E: Determinism-Preserving GPU Spatial Sharing for LLM Systems Workshop Paper
Zhenyuan Yang (ISCAS / UCAS), Wenxin Zheng (Shanghai Jiao Tong University), Mingyu Li (ISCAS / UCAS)
11:10–11:25
ESPEON: Verifying the Predestined Execution of Deterministic LLM Inference Workshop Paper
You-Liang Huang (Boston University), Yi Pan (UC Berkeley), Cheng Tan (Northeastern University), Yigong Hu (Boston University)
11:25–11:40
DriftBench: Measuring and Predicting Infrastructure Drift in LLM Serving Systems High-Impact Work (MLSys '26)
Gianluigi Vitale
11:40–11:45
TTTR: Record, Replay, and Steer AI Extended Abstract
Cheng Tan (Northeastern University)

11:45–13:30

Lunch

13:30–14:30

Keynote II

Hong Xu — Title TBD

Session Chair: TBD

14:30–15:30

Session 2: Reliability in Training & Serving

Session Chair: TBD

14:30–14:45
SDCs in the Wild: Characterizing and Diagnosing SDC-Defective GPUs in Production LLM Training High-Impact Work (OSDI '26)
Wenxin Zheng (Shanghai Jiao Tong University)
14:45–15:00
TTrace: Lightweight Error Checking and Diagnosis for Distributed Training High-Impact Work (NeurIPS '26 spotlight)
Haitian Jiang (New York University)
15:00–15:15
DuetScale: Scaling Disaggregated LLM Serving with Joint Parallelism and Startup Optimization Workshop Paper
Authors TBA
15:15–15:30
SemServe: A Systematic View of Exploiting Application Semantics in Agentic Serving Workshop Paper
Yaohua Xi, Minchen Yu (The Chinese University of Hong Kong, Shenzhen)

15:30–16:00

Coffee Break

16:00–16:25

Session 3: Correctness, Evaluation & Verification

Session Chair: TBD

16:00–16:15
High-Entropy String Hallucinations in LLM Inference: Quantization Sensitivity and Source-Constrained Decoding Workshop Paper
Jiwei Xu (Beijing University of Chemical Technology), Wei Jiang (The National Police University for Criminal Justice), Tao Wang (ISCAS)
16:15–16:20
SICD: A Domain-Aware Evaluation of Semantic Drift to Adversarial Diagnostic Interference Extended Abstract
Shouraya Pendgaonkar (Hightstown High School), Sameeksha Pendgaonkar (University of Pittsburgh)
16:20–16:25
Can Formal Verification Match AI Code Generation Speed? Extended Abstract
Mingyu Li, Zhanbo Wang (ISCAS / UCAS)

16:25–16:50

Open Discussion: What Does “Rigorous” Mean for LLM Systems?

Moderator TBD

16:50–17:00

Closing Remarks

Workshop Program