Three years go Zach N. Hofer-Shall and martin compston convinced me to join Twitter to build out the DevRel function to enable academic research with the #TwitterAPI It is the best decision I made. I will miss you, but the impact both of you had on me (and this program) is way bigger than you know

I'm noticing a trend.

The dictator is not amused. Funny.

In 2017, a team led by Andrew Ng published a paper showing off a Deep Learning model to detect pneumonia. Andrew is one of the most recognized researchers in the world, and the paper showed excellent results. But there was a big problem with their results: 1 of 8

Jetzt #zukunftsichern

Warnungen vor weitreichenden(!) negativen Konsequenzen für Studierende, Forschung und unseren Wissenschaftsstandort. Jetzt #zukunftsichern

I wondered whether it was just my bias that the home timeline content has become much worse over the past few days. But then Twitter provided the answer.

My take: search is a search problem, not a generation problem. Manifold interpolation is a good fit for generative tasks (like producing derivative poetry or images), but doesn't work for search (i.e. information retrieval).

Timothy B. Lee Just confirmed that I can’t tweet my sigmoid dot social profile

Looks like Twitter API for academic research is no more a thing. The link to apply for access seems broken.

Starting February 9, we will no longer support free access to the Twitter API, both v2 and v1.1. A paid basic tier will be available instead 🧵

The writing has been on the wall for some time now. I bet future access to the API will be reserved only for commercial purposes. I would be totally surprised if the free access to the full archive of Twitter for academic research doesn't get axed.

1/ In 2021, we shared next-gen language + conversation capabilities powered by our Language Model for Dialogue Applications (LaMDA). Coming soon: Bard, a new experimental conversational #GoogleAI service powered by LaMDA. blog.google/technology/ai/…

Bard is an experimental conversational AI service, powered by LaMDA. Built using our large language models and drawing on information from the web, it’s a launchpad for curiosity and can help simplify complex topics → goo.gle/3HBZQtu

Happy to announce that "User Privacy in Recommender Systems" has been accepted to the Doctoral Consortium of ECIR 2023 Dominik Kowald @elisab79 Know Center

We welcome contributions to 22 workshops at #TheWebConf 2023. Many workshops are still accepting contributions, and some have extended their deadlines. Check out the complete list at www2023.thewebconf.org/program/worksh. The conference will take place in Texas on April 30 - May 4 2023.

A broad survey of published methods to "augment" Language Models so they can reason, plan, and use tools to elaborate their answers. Tools such as search engines, calculators, code interpreters, database queries, etc, can help LLMs produce factual answers. By @MetaAI - FAIR.

Experts have known for years that current (auto-regressive) LLMs are - incredible - create bullshit - can be useful - are actually stupid - aren't actually scary

Had an insightful conversation with Geoffrey Hinton about AI and catastrophic risks. Two thoughts we want to share: (i) It's important that AI scientists reach consensus on risks-similar to climate scientists, who have rough consensus on climate change-to shape good policy.

Looking forward to my presentation tomorrow 11:50 at the #IR4Good Day ECIR2024