News & trends in the Last 48 hours
News & trends in the Last 48 hours
In the Last 48 hours there's been 14 influencers like MSFTResearch and peteskomoroch, discussing topics such as #AIforGood, #DeepLearningand #GreenAI.
Trends
Top hashtags
Top influencers
MSFTResearch 
peteskomoroch 
RichardSocher 
WIREDScience 
hardmaru 
stanfordnlp 
xamat 
yoavgo 
OpenAI Top sources
News
CS224n: Natural Language Processing with Deep Learning
RT @ImAI_Eruel: スタンフォード大学の深層学習を使った自然言語処理に関する人気講義"Natural Language Processing with Deep Learning"が開講され，講義の資料や,推薦図書・文献が公開されています
https://t.co/tTWXY1zrtf
言語処理の講義では間違いなく世界最先端で,内容は近年のNLP研究の動向をかなり反映してます https://t.co/bkOMcHHH7x
OpenCourse Description Natural language processing (NLP) is one of the most important technologies of the information age. There are a large variety of underlying tasks and machine learning models behind NLP applications. On the model side we will cover word vector representations, ...
RT @ImAI_Eruel: スタンフォード大学の深層学習を使った自然言語処理に関する人気講義"Natural Language Processing with Deep Learning"が開講され，講義の資料や,推薦図書・文献が公開されています
https://t.co/tTWXY1zrtf
言語処理の講義では間違いなく世界最先端で,内容は近年のNLP研究の動向をかなり反映してます https://t.co/bkOMcHHH7x
OpenCS224n: Natural Language Processing with Deep Learning
Course Description Natural language processing (NLP) is one of the most important technologies of the information age. There are a large variety of underlying tasks and machine learning ...
RT @SchmidhuberAI: 30-year anniversary of Very #DeepLearning (1991). Unsupervised hierarchical #PredictiveCoding finds compact internal representations to facilitate downstream learning. 1993: solving problems of depth 1000. Hierarchy can be distilled into a single deep net https://t.co/IQ5s8ZQQir
Open1991: First very deep learning with unsupervised pre-training
It uses unsupervised learning and predictive coding in a deep hierarchy of recurrent neural networks (RNNs) to find compact internal representations of long sequences of data. In other ...
RT @ChrSzegedy: Ongoing seminar "New Technologies in Mathematics: at Harvard:
https://t.co/Kz7JNjGgc3
With links to Video/slides of Josef Urban on AI for theorem proving to my talk on Language Modeling for mathematical reasoning.
OpenNew Technologies in Mathematics Seminar Series
of Computer Science, Princeton University Title: A Mathematical Exploration of Why Language Models Help Solve Downstream Tasks Abstract: Autoregressive language models pretrained on large ...
RT @tw_killian: @xamat There's been a few attempts to do just this in RL over the last few years.
First, Deepmind: https://t.co/NS3TV4cwpk
Then Berkeley/OpenAI: https://t.co/UP76gqsBXb
And then there was a flavor of this w/ Neural Episodic Control that didn't fully catch on: https://t.co/owztEXt8Nj
Open: Fast Reinforcement Learning via Slow Reinforcement Learning
Deep reinforcement learning (deep RL) has been successful in learning sophisticated behaviors automatically; however, the learning process requires a huge number of trials. In contrast, ...
RT @SFResearch: Thank you to everyone who submitted a proposal to our third annual Salesforce AI Research Grant. We’re proud to announce our 2020 round of winners. Congratulations!!
@bluevincent @Diyi_Yang @mutembesa @danqi_chen
Read More: https://t.co/wiTV9lbsMe
OpenCelebrating the Winners of the Third Annual Salesforce AI Research Grant
We are proud to announce the 2020 winners of our Salesforce AI Research Grant. Each of our winners will receive a $50K grant to advance their work and help us shape the future of AI.
RT @ashleyllorens: Today I start my first full week leading the global outreach team @MSFTResearch. Extremely grateful for this opportunity to help amplify the impact of research @Microsoft and advance the cause of science and technology research around the world: https://t.co/dR65VkikQJ
OpenAshley Llorens
I lead the global outreach team at Microsoft Research (MSR) – our mission is to amplify the impact of research at Microsoft and to advance the…
Machine learning programs trained with patient reports, rather than doctors', find problems that doctors miss—especially in Black people. https://t.co/u967lmupmV
OpenNew Algorithms Could Reduce Racial Disparities in Health Care
Machine learning programs trained with patient reports, rather than doctors', find problems that doctors miss—especially in Black people.
DALL-E drew laughs for creating images of a daikon radish in a tutu. But it builds on an important advance in computer vision with serious applications. https://t.co/OaM2gwwPO0
OpenThis AI Could Go From ‘Art’ to Steering a Self-Driving Car
DALL-E drew laughs for creating images of a daikon radish in a tutu. But it builds on an important advance in computer vision with serious applications.
RT @lintool: @redpony Aside: for the audience lacking context, @redpony and I, literally, wrote a book about scale in NLP together: https://t.co/l6S5LDEwRs
OpenData-Intensive Text Processing with MapReduce
Processing the enormous quantities of data necessary for these advances requires large clusters, making distributed computing paradigms more crucial than ever. MapReduce is a programming ...
@emilymbender Here is an example of a (very cool) application of BERT in a corpus study that explores atypical animacy.
https://t.co/0nlJRPMnnY
OpenClick here to read the article
Conceptualization Mariona Coll Ardanuy1,5 Daniel CS Wilson1,5 Methodology Mariona Coll Ardanuy Federico Nanni1 Kasra Hosseini1 Implementation Federico Nanni Kasra Hosseini Mariona Coll ...
Browse topics
Get updates live! Follow us on
Congratulations to the 2021 Microsoft Research Ada Lovelace Fellowship recipients! The program aims to increase the pipeline of diverse talent receiving advanced degrees in computing-related fields.
Learn more about our third cohort of fellows: https://t.co/7C6t2Ms7gM
OpenAda Lovelace Fellowship
Ada Lovelace Fellowship 2019: Nominations being accepted now through October 9, 2018
Webinar, tomorrow at 17:00 Paris time / 11:00 AM EST!
With the Prodigious Martial Hébert (Dean of CMU School of Computing).
1st in a series in which speakers are expatriated French AI researchers.
I'll be talking on March 3.
[Talks are in English].
https://t.co/AgYIBm289v
OpenFrance is AI Talks: monthly events by the PRAIRIE Institute with France Digitale | Inria
France Digitale and the PRAIRIE Institute are launching France is AI Talks, a monthly event with the best French AI researchers living abroad. The year’s programme will begin with Yann Le ...
Microsoft Excel the programming language is evolving. With the recent release of LAMBDA, users can now define new functions in the program’s formula language. Learn about the capability & its role in the transformation of spreadsheet formulas: https://t.co/gl17HXESks #POPL2021
OpenLAMBDA: The ultimate Excel worksheet function
Microsoft Excel the programming language is evolving. With the recent release of LAMBDA, users can now define new functions in the program’s formula language. Learn about the capability and ...
Diagnosing breast cancer is challenging but determining the best therapy is even harder. ReceptorNet from @SFResearch helps doctors make clinical decisions for breast cancer treatment therapy via cues not visible to the human eye. https://t.co/4d6isqTNYQ by @nikhil_ai et al.
OpenReceptorNet Helps Doctors Make Clinical Decisions for Breast Cancer Treatment
Part of the difficulty of treating breast cancer is the genomic variability of the patient population, as well as the different forms the cancer can take.
RT @allen_ai: The TLDR feature on @SemanticScholar is a single-sentence summary that helps you get the gist of a paper quickly, right on the search results page – check out this TLDR for our "Green AI" paper.
Want to give it a read? Find it here:
https://t.co/83YIkPbJV5
#GreenAI #AIforGood https://t.co/MeOL2zcRuV
OpenGreen AI
Creating efficiency in AI research will decrease its carbon footprint and increase its inclusivity as deep learning study should not require the deepest pockets.
RT @susansivek: Fascinating research on how descriptions of chatbots and metaphors employed can shape user expectations and evaluation. How to Build a Likable Chatbot https://t.co/bTnrlAMCRF https://t.co/PAbH3geAsX
OpenHow to Build a Likable Chatbot
Consumers have consistent personality preferences for their online friends, new research shows.
We're excited to announce the 2021 Microsoft Research PhD Fellowship recipients! For over two decades, our PhD Fellowship has supported nearly 300 outstanding students pursuing exceptional research.
Learn more about our 10 new fellows: https://t.co/7Ogsk9jc0K
OpenPhD Fellowship
The Microsoft PhD Fellowship Program is a two-year fellowship for PhD students enrolled in Computer Science, Electrical Engineering, or Mathematics.
Computation influences geopolitics (chip supply chain), IT sovereignty (GAIA-X), and qs about accessibility/efficiency of 'AI infrastructure'. So it's weird governments don't track it. That's going to change via a @OECD project I'm proud to be a part of! https://t.co/CYKkg6hHn3 https://t.co/rhyUyRxrfu
OpenWhy the OECD wants to calculate the AI compute needs of national governments
The OECD is working to help countries better understand their AI compute demand needs and prepare for the future.
RT @yoavgo: an amusing observation:
today's so called "huge LMs" are only very recently catching up in param size, and are still way WAAY behind in training data size, compared to the ngram-based LMs used in Google Translate back in 2007.
https://t.co/HwmRi0brPb
OpenLarge Language Models in Machine Translation
Simplified, the map, sharding and reduce functions do the following: Map(string key, string value) { // key=docid, ignored; value=document array words = Tokenize(value); hash_map histo; for ...
Scaling Kubernetes to 7,500 Nodes
We've scaled Kubernetes clusters to 7,500 nodes, producing a scalable infrastructure for large models like GPT-3, CLIP, and DALL·E, but also for rapid small-scale iterative research such as ...
RT @ACMRecSys: Are you an industry practitioner with a story tell? Then this one is for you !
The #RecSys2021 Call for Industry Talk Proposal is out. Check out our website for this, and of course other calls:
https://t.co/iagx2wMGaH
OpenCall for Contributions
Preference elicitation Privacy and security Socially- and context-aware recommender systems Systems challenges such as scalability, data quality, and performance User studies Authors will ...