ΑΙhub 前天 16:00
AIhub monthly digest: April 2025 – aligning GenAI with technical standards, ML applied to semiconductor manufacturing, and social choice problems
index_new5.html
../../../zaker_core/zaker_tpl_static/wap/tpl_guoji1.html

 

本月AI月报聚焦多个前沿领域,包括生成式AI的技术标准对齐、机器学习在半导体制造中的应用、社会选择问题的研究,以及AI Song Contest的回归。此外,还介绍了关于对抗攻击下机器学习模型的安全认证、数字时代工作方式的视觉再现比赛,以及英国的智能机器战略。报告还探讨了公众对AI的态度,并提供了相关资源和活动信息。

💡 Joseph Marvin Imperial的研究重点在于**将生成式AI与技术标准对齐**,以确保其符合法规和运营要求,这对于AI的合规性和安全性至关重要。

🔬 Amina Mević的研究涉及机器学习、物理学、数学和半导体技术的交叉领域,她专注于**将机器学习应用于半导体制造**,并结合了心理学和伦理学的元素,探索该领域的前沿发展。

🛡️ Chen Feng及其团队在AAAI 2025上发表了关于**对抗攻击下机器学习模型的安全认证**的研究,通过PROSAC方法,为机器学习模型在面对对抗攻击时提供了性能保障,并给出了人口水平的风险保证。

🤝 Eden Hartman团队的研究着眼于**解决社会选择问题**,他们提出了将Leximin公平性问题转化为功利主义优化的方案,为解决社会公平性问题提供了新的思路。

🎤 AI Song Contest将于2025年回归,这是一项国际比赛,旨在展示**人类与AI在歌曲创作过程中的共同创造潜力**,鼓励创新和合作。

Welcome to our monthly digest, where you can catch up with any AIhub stories you may have missed, peruse the latest news, recap recent events, and more. This month, we find out about aligning generative AI with technical standards, learn how machine learning can be applied to semiconductor manufacturing, investigate social choice problems, and hear about the return of the AI Song Contest.

Interview with Joseph Marvin Imperial: Aligning generative AI with technical standards

We continued our series meeting the AAAI/SIGAI Doctoral Consortium participants in this interview with Joseph Marvin Imperial. Joseph is based at the University of Bath, focusing on aligning generative AI with technical standards for regulatory and operational compliance.

Interview with Amina Mević: Machine learning applied to semiconductor manufacturing

Amina Mević is another AAAI/SIGAI Doctoral Consortium participant, working at the intersection of machine learning, physics, mathematics, and semiconductor technology. Her research also incorporates elements of psychology and ethics. She told us about her PhD research so far, what makes this field so interesting, and how she found the AAAI Doctoral Consortium experience.

Provably safe certification for machine learning models under adversarial attacks

In their work PROSAC: Provably Safe Certification for Machine Learning Models under Adversarial Attacks presented at AAAI 2025, Chen Feng, Ziquan Liu, Zhuo Zhi, Ilija Bogunovic, Carsten Gerner-Beuerle, and Miguel Rodrigues developed a new way to certify the performance of machine learning models in the presence of adversarial attacks with population-level risk guarantees. Here, Chen tells us more about their methodology, the main findings, and some of the implications of this work.

Investigating social choice problems

In their paper Reducing Leximin Fairness to Utilitarian Optimization, Eden Hartman, Yonatan Aumann, Avinatan Hassidim and Erel Segal-Halevi present a scheme for addressing social choice problems. In this interview, Eden tells us more about such problems, the team’s methodology, and why this is such a fascinating and challenging area for study.

AAAI 2025 workshops round-up

This month, we heard from the organisers of two AAAI 2025 workshops. Peter Santhanam filled us in on the key takeaways from the workshop on open source AI for mainstream use, whilst Daniela Annunziata and Marzia Canzaniello brought us up-to-speed with the latest happenings in the world of federated learning for unbounded and intelligent decentralization.

Open competition to create more realistic stock images of “digital transformation at work”

The ESRC Digital Futures at Work Research Centre and Better Images of AI have announced a competition to reimagine the visual communication of how work is changing in the digital age. There are eight prizes on offer for creation of new visual images across four themes: digital adoption, digital inclusion, changing employment contracts and working conditions, and digital dialogues. Artists and makers have until 18th May 2025 to enter. Find out more about the competition, and how to enter, here.

UK smart machines strategy 2035

The UK’s Robotics Growth Partnership (an independent expert committee appointed by the UK government) has published a roadmap to position the UK as a global leader in robotics. The report emphasizes the “transformative potential of smart machines to address pressing societal challenges, enhance economic productivity, and establish national leadership in a rapidly evolving technological landscape”.

How do people feel about AI?

A survey published by the Ada Lovelace Institute and the Alan Turing Institute reveals attitudes towards AI among the general public in the UK. It follows a previous survey carried out in 2022, before the release of ChatGPT and other LLM-based chatbots, which was published in 2023. The survey found that public awareness of different AI uses varies widely. While 93% have heard of driverless cars and 90% of facial recognition in policing, only 18% were aware of the use of AI for welfare benefits assessments. You can read more about the key findings here.

AI Song Contest back for 2025

The AI Song Contest is set to return in 2025, with the Award Show to take place in November in Amsterdam. The Contest is an international competition designed to showcase the creative potential of human–AI co-creativity in the songwriting process. If you are interested in taking part, you can find out more here.


Our resources page
Our events page
Seminars in 2025
AAAI/ACM SIGAI Doctoral Consortium interview series
AAAI Fellows interview series
AfriClimate AI series
AI around the world focus series

Fish AI Reader

Fish AI Reader

AI辅助创作,多种专业模板,深度分析,高质量内容生成。从观点提取到深度思考,FishAI为您提供全方位的创作支持。新版本引入自定义参数,让您的创作更加个性化和精准。

FishAI

FishAI

鱼阅,AI 时代的下一个智能信息助手,助你摆脱信息焦虑

联系邮箱 441953276@qq.com

相关标签

生成式AI 机器学习 半导体制造 社会选择 AI Song Contest
相关文章