Thanks so much to @PVanHentenryck for hosting me at @GeorgiaTech
's @Ai4Opt
, one of their three @AIINSTITUTES
. I gave a talk on efficient private mean estimation, a #STOC2022 paper with @Samuel_BKH @mahbodm_
, check out the recording: https://
youtube.com/watch?v=gXmijL
3o8SI
…
RESEARCH
-
Efficient Private Mean Estimation Research Presented at Georgia Tech
By
–
-
Deep Learning Models: Improving Robustness and Efficiency Through Algorithms Research
By
–
As #DeepLearning models become more widely used, it is increasingly important that they be both robust and efficient. Today we summarize some of our many efforts to improve #ML efficiency through algorithms research. → https://
goo.gle/3I6asCj -
E3B: New Method for Exploring Complex Variable Environments
By
–
E3B is a method for exploring complex environments which vary across episodes. This work set a new SOTA for MiniHack & reward-free exploration on Habitat in October.
— AI at Meta (@AIatMeta) 7 février 2023
Read the paper ➡️ https://t.co/5oxIf1kDAy
Get the code ➡️ https://t.co/omz9GLkkha pic.twitter.com/elUIs4FIofE3B is a method for exploring complex environments which vary across episodes. This work set a new SOTA for MiniHack & reward-free exploration on Habitat in October. Read the paper https://
bit.ly/3JUCceJ
Get the code https://
bit.ly/3YulsPk -
Microsoft and Google conferences: AI announcements reactions
By
–
Mañana hacemos directo viendo y reaccionando a las conferencias de Microsoft (hoy) y la de Google (mañana). Será en diferido y dependerá de si hay sustancia sumando lo anunciado en ambas. De Microsoft ya sabemos esto
-
MIT Researchers Discover LLMs Store Neural Networks Internally
By
–
MIT researchers found that massive neural nets (e.g. large language models) are capable of storing and simulating other neural networks inside their hidden layers, which enables LLM to adapt to a new task without external training: https://
bit.ly/3jIiesP -
Adding Example Rows Improves SQL Database Chain Performance
By
–
SQLDatabase chain # of rows Research by Nitarshan Rajkumar, Raymond Li, Dzmitry Bahdanau shows that the more rows you add into the prompt, the better you can query SQL s/o to Jon Luo for adding Paper: https://
arxiv.org/abs/2204.00498 Example: https://
langchain.readthedocs.io/en/latest/modu
les/chains/examples/sqlite.html#adding-example-rows-from-each-table
… -
Cerebras Wafer-Scale Engine Runs First CFD Simulation
By
–
Today, with @NETL_DOE & @psc_live we announced the first ever computational fluid dynamics simulation on our Wafer-Scale Engine. Running on @NSF-funded Neocortex, this natural convection simulation promises more insightful #scientificcomputing https://t.co/ud3sL53WZg pic.twitter.com/zCDvKISb8H
— Cerebras (@cerebras) 7 février 2023Today, with @NETL_DOE & @psc_live we announced the first ever computational fluid dynamics simulation on our Wafer-Scale Engine. Running on @NSF
-funded Neocortex, this natural convection simulation promises more insightful #scientificcomputing http://
cerebras.net/cfd -
Neural Attention Mechanism: The Key to Transformer Success
By
–
The “neural attention mechanism” is the secret sauce that makes transformers successful on various tasks and datasets. Check out our latest blog post on vision transformers (ViTs).https://t.co/MvPPYKZ5nC#ai #computervision #visiontransformer pic.twitter.com/psxhbtaeYL
— Satya Mallick (@LearnOpenCV) 7 février 2023The “neural attention mechanism” is the secret sauce that makes transformers successful on various tasks and datasets. Check out our latest blog post on vision transformers (ViTs). https://
learnopencv.com/attention-mech
anism-in-transformer-neural-networks/
… #ai #computervision #visiontransformer -
Google LaMDA Becomes Main Project to Counter ChatGPT
By
–
The same Google LaMDA that had everyone going crazy last summer over claims of sentience is now the company’s main project to stay relevant in a ChatGPT age. // $GOOGL @OpenAI @GoogleAI // #AI #artificialintelligence #deeplearning #chatgpt #LaMDA
-
Prophecy Markets: The Future of AI Prediction and Investment
By
–
Today's essay is a riff on @js_horne
's Prophecy Markets, one of my favorite ideas in a long time.