Yawen Duan
@yawen_duan
Followers
348
Following
186
Media
17
Statuses
130
Concordia AI https://t.co/Pe2BhjbbE0 | Frontier AI Safety & Governance
Beijing, China
Joined January 2019
🚨 New 150+ page report on the "State of AI Safety in China" by Concordia AI  🇨🇳. Learn about China’s role in tackling frontier AI risks through Concordia AI’s new report.
1
6
23
RT @DanHendrycks: We’re releasing Humanity’s Last Exam, a dataset with 3,000 questions developed with hundreds of subject matter experts to….
0
773
0
RT @_achan96_: Whether 2025/6/7/8/etc is the year of AI agents, we will need tools to help us unlock their benefits and manage their risks.….
0
30
0
RT @haonanlp: We are excited to announce Libra-Leaderboard: The first LLM leaderboard dedicated to balancing safety and capability in LLMs.….
0
15
0
RT @valentina__py: 📣 We are looking forward to an interesting line up of invited speakers at the SoLaR workshop at NeurIPS! . Come join us….
0
16
0
RT @mattsheehan88: Back in 2017 I started wondering what 🇨🇳 scientists & policymakers thought about AI safety. I asked around, didn’t find….
carnegieendowment.org
Beijing’s AI safety concerns are higher on the priority list, but they remain tied up in geopolitical competition and technological advancement.
0
40
0
RT @solarneurips: We're soliciting reviewers for SoLaR! We expect a lightweight assignment for all reviewers (2-3 papers max) and welcome r….
docs.google.com
The workshop on Socially Responsible Language Modeling Research (SoLaR) 2.0 will be at NeurIPS 2024 and is organized, among others, by David Krueger, Maarten Sap and Yejin Choi. We will bring...
0
6
0
Excited to announce the next iteration of SoLaR workshop to appear on @NeurIPSConf 2024. We focus on Socially Responsible Language Modelling Research!.Follow our official account, and stay tuned for more updates!.
We are thrilled to announce that there will be another edition of Socially Responsible Language Modelling Research (SoLaR) workshop at NeurIPS2024! This year we will host very important speakers, and we're opening to submissions soon. Stay tuned for more future announcements!
0
0
10
fwiw, I started looking into the taxonomy/framework in the paper (esp. swiss cheese model and risk decomposition) since I was a facilitator of early cohorts of MLSafety Course in 2022, and has been one of my top lenses to view AI safety research works. Highly recommended!.
A retrospective of Unsolved Problems in ML Safety. Unsolved Problems, written the summer of 2021, mentions ideas that were nascent or novel for their time. Here are a few:. • Hazardous Capabilities Evals: In the monitoring section, we introduce the idea.
0
0
3
RT @farairesearch: ICYMI: Here’s highlights from our previous research on "Adversarial Policies Beat Superhuman Go AIs." We found that even….
0
5
0
RT @TiFA_ICML2024: A kind reminder that the submission of our workshop is about to close in ~1 day. 🌏 Workshop Website: .
openreview.net
Welcome to the OpenReview homepage for ICML 2024 Workshop TiFA
0
3
0
We're rly grateful for @HaydnBelfield's kind words. ---.Check our website ( for the latest "State of AI Safety in China" Spring 24' report. Also, webinar ( w/ guests @jjding99, @AngelaZhangHK , @mattsheehan88, and @RobertTrager.
This is the best single report for understanding the Chinese AI Governance landscape. Concordia have made valuable contributions to the Bletchley & Seoul Summits and Bengio's IPCC-style Report
0
7
17
RT @AISafetyInst: We are announcing new grants for research into systemic AI safety. Initially backed by up to £8.5 million, this program….
0
80
0
big.
AI agents, which could accomplish complex tasks with limited human supervision, are coming down the pipe. How do we manage their risks?. Our new @FAccTConference paper argues that we need visibility---information about the use of agents---and investigates how to obtain it. 🧵
0
0
2
RT @yi_zeng: It is always great to have you in Beijing. The discussion is about near term and long terms risks and challenges of and from A….
0
3
0
RT @S_OhEigeartaigh: One of the highlights of my China trip was getting to discuss AI alignment on one of China's leading TV chat shows, al….
0
6
0
🚀 Excited to announce the Trustworthy Multimodal Foundation Models and AI Agents (TiFA) workshop @ICML2024, Vienna!. Submission deadline: May 30 EOD AoE.
icml-tifa.github.io
ICML TiFA WorkShop
🚀 Excited to announce the Trustworthy Multimodal Foundation Models and AI Agents (TiFA) workshop @ICML2024, Vienna! Submission deadline: May 30 EOD AoE.
0
2
10
RT @DavidSKrueger: I’m super excited to release our 100+ page collaborative agenda - led by @usmananwar391 - on “Foundational Challenges In….
0
154
0
RT @ancadianadragan: RS and RE roles, growing our bay area presence as part of our further investment in safety and alignment:.
0
7
0