Global AI Weekly

Issue number: 74 | Tuesday, October 15, 2024

Highlights

Demis Hassabis: from video game designer to Nobel prize winner

Demis Hassabis: from video game designer to Nobel prize winner

Google DeepMind chief believes in benefits of AI but says risks must be taken as seriously as the climate crisis Most 17-year-olds spend their days playing video games, but Britain’s latest Nobel prize winner spent his teenage years developing them. Sir Demis Hassabis, who was jointly awarded the chemistry prize on Wednesday, got his big break in the tech world as co-designer of 1994’s hit game Theme Park, where players create and operate amusement parks.

theguardian.com

Someone claims to have used AI to apply to 2,843 jobs

Someone claims to have used AI to apply to 2,843 jobs

AI is spamming up job applications. Jason Koebler from 404 Media writes about a person who claimed to have used a free tool, AI Hawk, to apply for 17 jobs in an hour — only stopping when they’d reached 2,843. The tool automatically entered the person’s bio, generated résumés, wrote customized cover letters, and checked […]

techcrunch.com

MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering

We introduce MLE-bench, a benchmark for measuring how well AI agents perform at machine learning engineering.

openai.com

Video

Can OpenAI's ChatGPT Beat Humans At Gaming?

Can OpenAI's ChatGPT Beat Humans At Gaming?

If you wish to appear here or pick up other perks, click here: https://www.patreon.com/TwoMinutePapers My paper on simulations that look almost like reality is available for free here:

youtube.com

Articles

The promise and perils of synthetic data

The promise and perils of synthetic data

Is it possible for an AI to be trained just on data generated by another AI? It might sound like a harebrained idea. But it’s one that’s been around for quite some time — and as new, real data is increasingly hard to come by, it’s been gaining traction. Anthropic used some synthetic data to […]

techcrunch.com

Research Focus: Week of October 7, 2024

Research Focus: Week of October 7, 2024

Simplifying secure decision tree training; Improving accuracy of audio content detection; A novel neurosymbolic system for converting text to tables; New video series: AI for Business Transformation; TEE security protections for container workloads. The post Research Focus: Week of October 7, 2024 appeared first on Microsoft Research.

microsoft.com

LLM vs LLM: Codenames Tournament

LLM vs LLM: Codenames Tournament

A mini multi-agent competition among 3 different LLM agents.

towardsdatascience.com

ChatGPT will happily write you a thinly disguised horoscope

ChatGPT will happily write you a thinly disguised horoscope

There's a meme floating around at the moment where you ask ChatGPT the following and it appears to offer deep insight into your personality.

simonwillison.net

Florence-2: Mastering Multiple Vision Tasks with a Single VLM Model

Florence-2: Mastering Multiple Vision Tasks with a Single VLM Model

In this tutorial we introduce Florence-2 [1]— a novel, open-source Vision-Language Model (VLM) designed to handle a diverse range of vision and multimodal tasks, including captioning, object detection, segmentation and OCR. After installing and importing the necessary libraries (as demonstrated in the accompanying Colab notebook), we begin by loading the Florence-2 model, processor and the input image of a camera:

towardsdatascience.com

PyTorch Foundation Technical Advisory Council Elects New Leadership

PyTorch Foundation Technical Advisory Council Elects New Leadership

We are pleased to announce the first-ever Chair and Vice Chair of the PyTorch Foundation’s Technical Advisory Council (TAC): Luca Antiga as the Chair and Jiong Gong as Vice Chair.

pytorch.org

MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpainting

We introduce MaskedMimic a single unified controller for physically simulated humanoids. Our system is capable of generating a wide range of motions across diverse terrains from intuitive user-defined intents. In this work, we show several applications, including generating full-body motion from partial joint target positions, responding to joystick steering, engaging in object interactions, following paths, interpreting text commands, and even combining these modalities, such as executing text-stylized path following.

research.nvidia.com

New 3D printing technique creates unique objects quickly and with less waste

New 3D printing technique creates unique objects quickly and with less waste

By using a 3D printer like an iron, researchers can precisely control the color, shade, and texture of fabricated objects, using only one material.

news.mit.edu


Upcoming Events

Make Azure AI Real is Back - Join Microsoft AI Experts for Season 2 with the Reactor!

Make Azure AI Real is Back - Join Microsoft AI Experts for Season 2 with the Reactor!

Make Azure AI Real is a YouTube livestream series brought to you by Microsoft Reactor, where we don't just talk AI—we do AI! Register..

techcommunity.microsoft.com

Your home for Microsoft AI Tour

Your home for Microsoft AI Tour

Register for the Microsoft AI Tour and discover how AI can help you solve your most significant challenges and create new opportunities for your organization.

aitour.microsoft.com

>