The AI Agent Index: the first public database to document information about currently deployed agentic AI systems.
aiagentindex.mit.edu
@reshmighosh.bsky.social
Applied Scientist @Microsoft | PhD from @Carnegie Mellon | Working on Security, ethics and Trustworthy AI | IEEE Chair
The AI Agent Index: the first public database to document information about currently deployed agentic AI systems.
aiagentindex.mit.edu
Planet Mars looking orange and bright below centre. Star Pollux to its left with the star Castor above. Numerous other star dots on a dark background.
I've just submitted this picture of the triangle of Mars, Pollux, and Castor to Astronomy Now magazine. ๐ญ ๐งช ๐จ #astrophotography #SciArt #photography #StormHour #ThePhotoHour
31.01.2025 20:35 โ ๐ 655 ๐ 50 ๐ฌ 22 ๐ 1I think some people hear โgrantsโ and think that without them, scientists and government workers just have less stuff to play with at work. But grants fund salaries for students, academics, researchers, and people who work in all areas of public service.
โPausingโ grants means people donโt eat.
If youโre at a European institution and are looking to poach American scientists, this would be a very good week to do it
28.01.2025 08:31 โ ๐ 1183 ๐ 140 ๐ฌ 43 ๐ 14A low-res, slightly grainy mid 2000s camera phone photo of a white Wii stood next to a boxy silver crt TV with a remote control next to it
Wii love you ๐
28.01.2025 13:26 โ ๐ 2312 ๐ 118 ๐ฌ 53 ๐ 12Hello! I canโt make it today - but will you be organizing more events? I would love to join the community
24.01.2025 22:06 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0AMOC is a large-scale ocean current system that transports warm surface water northward and cold, dense deep seated water southward. This circulation is driven by differences in water temperature and salinity, further regulating heat distribution, ocean nutrient cycling and atmospheric conditions ๐๐งช
04.01.2025 02:35 โ ๐ 60 ๐ 18 ๐ฌ 2 ๐ 4H1B is for all countries - I am not sure how you are targetting only individuals from one country? Also most Indians I know are in tech - I donโt see the correlation between the screenshot you posted (with H1b salaries of cooks) and them being Indians? How are you establishing that relationship?
28.12.2024 06:28 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0deciding between "actually taking some time off" and "work on personal projects and call it relaxing"
hmm i feel attacked
24.12.2024 20:27 โ ๐ 96 ๐ 12 ๐ฌ 5 ๐ 0๐ขcalling for interested paper reviewers
If you have published work in ML automation, and make systems more effective, and are interested in taking on reviewer load in the coming months - please fill up the form below (attached in comments). We will contact you based on the load and fit.
Phase Transition xkcd.com/3025
16.12.2024 20:01 โ ๐ 20691 ๐ 2529 ๐ฌ 186 ๐ 142Tomorrow @jakublucki.bsky.social will be presenting the BEST TECHNICAL PAPER at the SoLaR workshop at NeurIPS. Come check our poster and his oral presentation!
14.12.2024 03:43 โ ๐ 7 ๐ 1 ๐ฌ 0 ๐ 0Hey there. Here is a paper I wrote on autonomous agents that autonomously navigate the desktop user interfaces.
โTowards Automated Exploration of Interactive Systemsโ
Sorry itโs not on arXiv right now. Hereโs the PDF: faculty.cc.gatech.edu/~riedl/pubs/...
We've launched an official City of Boston starter pack to make it easy to follow City departments. This starter pack will stay up to date as more teams join the Bluesky community. Give them a follow!
12.12.2024 20:40 โ ๐ 421 ๐ 90 ๐ฌ 20 ๐ 41An eye for an eye, and we all go blind.
13.12.2024 18:40 โ ๐ 14220 ๐ 1853 ๐ฌ 879 ๐ 255Working on finding alignment gaps in the latest AI models? Do you have applications that suggest LLMs/MMLMs are doing better than human workers? Have you proved the need to work on aligning AI to humans?
Yes? If you are working on alignment research-consider submitting a paper to our workshop@ICLR
(3) A Survey on LLM-as-a-Judge
(4) Predicting Emergent Capabilities by Finetuning
(5) SAGEval: The frontiers of Satisfactory Agent based NLG Evaluation for reference-free open-ended text
(6) CS-Eval: A Comprehensive Large Language Model Benchmark for CyberSecurity
Hey Boston, weโre here!
City teams are building accounts and making moves to make Bluesky our home.
Stay tuned. In the meantime, you can follow a few of our first official accounts: @boston-streets.bsky.social, @bostonparks.bsky.social, and @healthyboston.bsky.social.
(1/4) The City is placing toy and gift donation collection boxes at City buildings, libraries, and firehouses from Tuesday, December 3, through Wednesday, December 18, at 4:30 p.m. Donations will be accepted at the following Somerville City buildings, libraries, and Firehouses:๐งต
03.12.2024 20:38 โ ๐ 5 ๐ 3 ๐ฌ 1 ๐ 1We could go on about how we welcome publishers, we don't demote links, we encourage independent developers to build apps and extensions on top of Bluesky's network.... but instead, we'll show you.
All thanks to the incredible community here! ๐ฆ
Iโm thankful for the top ten scientific developments that improve health:
Vaccines
Antibiotics & antimicrobials
Clean water, sanitation, & hygiene
Food safety & fortification innovations
Modern agriculture methods
Public health education campaigns
1/
You know how itโs awfully hard to verify the accuracy & correctness of outputs in RAG scenarios? Want to work as a PhD research intern to create & test different UIs that make this easier? Contact me.
26.11.2024 17:02 โ ๐ 4 ๐ 1 ๐ฌ 3 ๐ 0Interviewer: what's your greatest strength?
Me: i have exceptional hindsight
Interviewer: that doesn't really help us
Me: yes, i see that now
Runway just released a new foundation generative #AI image model โ called Frames
Allows you to define styles (called โWorldsโ) for consistency across multiple image generations
Adobe's DynaSaur
An LLM agent framework that model each action as a Python function. At each step, the agent generates Python code snippets. The generated code is executed through a Python interpreter, and the resulting observations are returned to the agent.
No one can explain stochastic gradient descent better than this panda.
24.11.2024 15:04 โ ๐ 218 ๐ 33 ๐ฌ 10 ๐ 6No you are not - everyone is using the same tools, the language included looks same..
Also easier said than done, claim the small wins you are shy of showing. It distinguishes you from your friends who are writing about the same clubs that everyone is part ofโฆ (n/n)