We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode AI for COVID detection fails, GitHub's Copilot can code, GAN Theft Auto is fun

AI for COVID detection fails, GitHub's Copilot can code, GAN Theft Auto is fun

2021/7/1
logo of podcast Last Week in AI

Last Week in AI
AI Deep Dive AI Chapters Transcript
People
A
Abigail See
A
Andrey Krenkov
D
Daniel Bashir
S
Sharon Zhou
Topics
Andrey Krenkov: 本期节目讨论了Google关于深度学习模型优化的综述论文，该论文探讨了如何使模型更小、更快、更好，并提出了一些经验证的优化策略。此外，还讨论了大量用于检测COVID-19的机器学习模型并不适合临床使用，研究人员发现这些模型存在问题，例如依赖图像中的无关信息，并提出改进建议，包括提高可解释性、改进数据收集和临床医生参与等。最后，还介绍了GitHub和OpenAI合作推出的代码生成工具Copilot，以及亚马逊仓库机器人技术、Twitter的AI伦理团队、LinkedIn算法偏差等话题。 Sharon Zhou: 就模型优化而言，她认为应根据模型对资源和质量的敏感度选择不同的方法：资源敏感型选择高效架构，质量敏感型则先训练大型模型再压缩。在讨论AI医疗诊断的局限性时，她强调了模型可解释性的重要性以及临床医生参与研究的重要性。 Abigail See: 她分享了自己在使用GPT模型方面的经验，并对Copilot代码生成工具的潜力和局限性进行了分析，指出该工具可能生成不安全或有偏见的代码。在讨论YouTube推荐算法时，她提到了视频病毒式传播的现象，并分析了其背后的机制。 Daniel Bashir: 他补充介绍了其他一些AI相关的新闻，包括Google将AI研究部门的员工转移到新的机器学习团队，丰田研究院在机器人技术方面的进展，现代汽车集团收购波士顿动力公司，DeepMind研究员呼吁对AI伦理负起集体责任，以及斯坦福大学启动的AI伦理审查项目等。

Deep Dive

Chapters
The discussion covers a survey on optimizing deep learning models for efficiency, focusing on techniques to make them smaller, faster, and better. The conversation explores the balance between model size and quality, and the practical implications for developers and researchers.

Shownotes Transcript

Our 62nd Last Week in AI episode with a summary and discussion of last week's big AI news! On this episode, more controllable GAN, more controllable language models, AI for wildfires and reading lips, being critical of human oversight of AI, surveillance by AI in China, and some fun AI weirdness.

This week:

-

Google Survey Explores Methods for Making DL Models ‘Smaller, Faster, and Better’)

-

Machine learning models that detect COVID-19 on chest X-rays are not suitable for clinical use)

-

GitHub and OpenAI launch a new AI tool that generates its own code)

-

What’s Going on With Amazon’s “High-Tech” Warehouse Robots?)

-

How Twitter hired tech's biggest critics to build ethical AI)

-

LinkedIn’s job-matching AI was biased. The company’s solution? More AI.)

-

GAN Theft Auto is a snippet of GTA 5 made by AI)

Find this and more in our text version of this news roundup:  https://lastweekin.ai/p/122

Music: Deliberate Thought, Inspired by Kevin MacLeod (incompetech.com)