Auto-formalization & Logical Reasoning
Turning natural-language arguments into executable first-order logic, then letting a solver do the reasoning. Drafting many formalizations, pruning the contradictory ones, and keeping what verifies.
PhD @ UC Berkeley EECS · Research Scientist Intern @ Meta
I teach neural networks to reason with logic — and make them harder to fool.
Hi! I'm a CS PhD student at UC Berkeley, advised by Prof. Pierluigi Nuzzo. I work where neural intuition meets symbolic rigor: teaching language models to reason in formal logic, so their conclusions are proven, not guessed. On the security side, I probe the safeguards we build around these models until they break.
Before Berkeley I spent two and a half years as a PhD student at the University of Southern California (2022–2024), and earned my B.S. in Physics from the University of Science and Technology of China as an Outstanding Graduate (top 3%). This year I'm a Research Scientist Intern at Meta.
If you're interested in reasoning, formal verification, or AI security, reach out — I'm always happy to collaborate.
Three threads, one question: can we trust what a model concludes?
Turning natural-language arguments into executable first-order logic, then letting a solver do the reasoning. Drafting many formalizations, pruning the contradictory ones, and keeping what verifies.
Agentic systems that come with something to check — assessor agents with structured failure tracking, and self-evolving contract frameworks for verified controller synthesis in cyber-physical systems.
Showing where defenses learn shortcuts instead of intent, building adversarially robust sparse GNNs, and secure cloud–device learning that keeps the informative features encrypted.
* equal contribution · ♦ equal advising · click a paper for highlights
An agentic framework where contracts evolve alongside the synthesized controller, so every generated controller comes with a specification it can be verified against.
Controlled diagnostics reveal position, token-trigger, and topic-generalization shortcuts in fine-tuned prompt-attack defenses. High rejection rates often come from surface correlations, not from understanding harmful intent.
An inference-time framework that drafts diverse plans, prunes contradictory or ambiguous formalizations, and aggregates verified paths, with no extra supervision.
A reproducible assessor-agent framework with structured failure tracking and standardized agent-to-agent interfaces, plus an auto-formalization agent that compiles natural-language logic into executable Z3Py under auditable budgets.
A secure cloud–device framework that combines hypernetwork-based personalization with selective feature encryption. It matches full fine-tuning without backpropagation on the device, and dynamic feature sampling encrypts only the informative channels.
GrokTransfer initializes a target model with embeddings learned by a weaker model. It provably generalizes directly on synthetic XOR and removes delayed generalization in MLPs and Transformers across tasks.
ARGS jointly prunes perturbed graph edges and GNN weights with label- and homophily-aware objectives. The result is highly sparse GNNs with much stronger adversarial robustness and lower inference cost.
Sunnyvale, CA · Meta AI
Building domain foundation models for ads ranking.
Berkeley, CA · Advisor: Prof. Pierluigi Nuzzo
Neuro-symbolic reasoning, auto-formalization, and verifiable agents.
Sunnyvale, CA · DSO.AI
Built an Analog Agent that identifies the function and structure of circuits automatically. A hybrid pipeline of GNNs and subgraph-isomorphism tests recovered 100% of non-overlapping blocks.
Los Angeles, CA · Electrical & Computer Engineering
Adversarial robustness of sparse graph neural networks (ICASSP 2024 Oral, TMLR 2024).
Hefei, China
Trained large multi-layer Transformers for bidirectional machine translation and beat Google Translate by 15% BLEU.
Remote · Prof. Junjie Zhu
Hefei, China · Outstanding Graduate (top 3%)
ACIES accepted to the REALM 2026 workshop, co-located with EMNLP 2026.
Draft-and-Prune is now on arXiv.
Agentified Assessment of Logical Reasoning Agents accepted at the ICLR 2026 AIWILD Workshop.
Joining Meta as a Research Scientist Intern in summer 2026.
Selected as a DAC 2025 Young Student Fellow.
Starting my first internship at Synopsys this summer.
Continuing my PhD at UC Berkeley EECS with Prof. Nuzzo.
Our first journal paper accepted to TMLR 2024.
Our first paper accepted as an oral at ICASSP 2024.
Admitted to the PhD program at USC.
Design Automation Conference
Adversarial vulnerabilities of graph lottery tickets
USTC · top 3%
Every glowing dot is a city someone has visited this page from. Thank you for stopping by.
Open to collaborations