So spiky science.org/doi/10.1126/sc…

It is very interesting to see how the current narrative of AI research today follows what the MIT Media Lab invented and pioneered almost 20 years ago. Pattie Maes research paper 'Agents that reduce work and information overload' was published in 1995: hiddenheroes.netguru.com/pattie-maes

Contrastive learning of representations can help discover quantitative imaging biomarkers for oncology!
📝 nature.com/articles/s4225…

☑️ Performant in low data
☑️ Resource-efficient
☑️ Robust to inter-reader and acquisition diff.

Nature Machine Intelligence AIM @ Harvard Brigham and Women’s Radiation Oncology
🧵1/6

Paper by Oxford Comp Sci researchers for EWADA Oxford Martin School in Nature Machine Intelligence highlights gaps in the application of ethics in AI for children and urges a more considered approach to ensure systems fully meet their needs.
Read more: tinyurl.com/48aaytxx

#compscioxford

No 'Zero-Shot' Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performance

Web-crawled pretraining datasets underlie the impressive 'zero-shot' evaluation performance of multimodal models, such as CLIP for classification/retrieval and Stable-

The generative AI tooling I'm most excited about augments creativity and does not replace it.

Pat Pataranutaporn explains by taking us back to the origins of the word cyborg.

From last year's AI on the Lot. MIT Media Lab

Book preview created by Seed

In the latest #ACMByteCast , in partnership w/@hanselminutes, Scott Hanselman 🌮 hosts ACM Fellow Rosalind Picard (@medialab), scientist, inventor, engineer, co-founder Empatica @affectiva.

Listen & subscribe: learning.acm.org/bytecast/ep51-…

Ehi Nosakhare Rob Morris ACM SIGCHI

Here's a close-up of hue and brightness heads in a vision transformer. Work done with tremendous help from Aoyu Wu and Cynthia Chen!

A simple implementation of the #life2vec model is now available at github.com/carlomarxdk/li…. You can try out the workflow with dummy data, + we are planning to add more examples (including finetuning) in the near future.

My data viz image was picked for a cover of the Jan 2024 issue of the Nat Comput Sci journal 😋

Many prior interpretability methods based on projections onto the vocab space (e.g., by Guy Dar Nora Belrose Jacob Steinhardt Evan Hernandez) and intervening on the LLM computation (e.g., by Kevin Meng Arthur Conmy Koyena Pal David Bau) can be viewed as 🩺instances. 2/9

🧵Can we “ask” an LLM to “translate” its own hidden representations into natural language? We propose 🩺Patchscopes, a new framework for decoding specific information from a representation by “patching” it into a separate inference pass, independently of its original context. 1/9

Contextualizing the idea of Cyborg and Human Augmentation in the context of Southeast Asia #SEAT2024

本日より21_21 DESIGN SIGHTにて開催される展覧会「未来のかけら: 科学とデザインの実験室」に稲見自在化身体プロジェクトと遠藤麻衣子監督が共作した短編映画『自在』の関連展示が出品されます。『自在』に使用される三つ目のメガネのデザインは今展覧会ディレクターの山中俊治さんによるものです。

Neuroscience of Consciousness

Where is the ghost in the shell?
academic.oup.com/nc/article/202…

さっきからのAIからIAへの話で、前のAIブームの10年前位も皆そんな話してたなーと思い出してたけど話繰り返すのね。このスライド８年前のなんかのシンポジウムでなるみさんが話してたやつ

Some predictions for 2024 – keeping only the more controversial ones. You certainly saw the non-controversial ones (multimodality, etc) already

1. At least 10 new unicorn companies building SOTA open foundation models in 2024
Stars are so aligned:
- a smart, small and dedicated…

.Thomas Wolf: “Academia is back as we saw at NeurIPS 2023. With many private and open-source labs closing the doors on publishing their results and data, academia rises again in visibility and is shining with many impactful papers in 2023 and exciting new work coming.”