Fwd: Jingxuan He on Nov 6

[Alan Karp]

--------------
Alan Karp

---------- Forwarded message ---------
From: Aditi Partap <adit...@stanford.edu>
Date: Fri, Oct 31, 2025 at 2:14 PM
Subject: Jingxuan He on Nov 6
To: security...@lists.stanford.edu <security...@lists.stanford.edu>

           Security: A Next Frontier in AI Coding

                        Jingxuan He

                 Thursday, November 6, 2025

                       Talk at 4:00pm

                      CoDA E201 & Zoom

Abstract:

AI is revolutionizing programming, but is it also creating a new

generation of security debt?  This talk confronts this critical question

through a comprehensive analysis of AI’s impact on the software security

landscape.  I will begin by introducing two novel benchmarks to quantify

the risks: CyberGym, to assess AI’s offensive capabilities in

vulnerability reproduction and discovery, and BaxBench, to measure AI’s

propensity to introduce security flaws when generating code.  Drawing on

insights from these evaluations, I will then present two proactive

approaches to make AI-generated code secure by design.  I will detail

our work in fine-tuning models on curated datasets of secure code

patterns to minimize the generation of vulnerabilities.  Furthermore, I

will describe an inference-time constraining mechanism that enforces

type safety in generated code, provably eliminating entire classes of

bugs.

Bio:

Jingxuan HE is a Postdoctoral Researcher at UC Berkeley, working with

Prof.  Dawn Song.  He completed his PhD at ETH Zurich, where he was

advised by Prof.  Martin Vechev.  His research lies at the intersection

of security, AI, and programming languages, with a current focus on

evaluating AI’s impact on cybersecurity and mitigating associated risks.

His work has been recognized with an ACM CCS Distinguished Paper Award

and an ETH Medal for Outstanding Doctoral Thesis, and is adopted for

model evaluations by leading AI labs like Anthropic.  For more

information, please visit https://jxhe.info/.

--++**==--++**==--++**==--++**==--++**==--++**==--++**==
security-seminar mailing list
security...@lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/security-seminar