Pre-Trained Language Models for Interactive Decision-Making
— AK (@ak92501) February 4, 2022
abs: https://t.co/uECv8kutrE
project page: https://t.co/Bf3iqgfcA9 pic.twitter.com/OLSIiOxX2S
Pre-Trained Language Models for Interactive Decision-Making
— AK (@ak92501) February 4, 2022
abs: https://t.co/uECv8kutrE
project page: https://t.co/Bf3iqgfcA9 pic.twitter.com/OLSIiOxX2S
Unified Scaling Laws for Routed Language Models
— AK (@ak92501) February 3, 2022
abs: https://t.co/C4zMJcB2wg pic.twitter.com/LoKuIVW617
Competition-Level Code Generation with AlphaCode
— AK (@ak92501) February 2, 2022
paper: https://t.co/Np8uy6UE3R
blog: https://t.co/ATpcgHNeGB pic.twitter.com/x3iGv5UjBM
WebFormer: The Web-page Transformer for Structure Information Extraction
— AK (@ak92501) February 2, 2022
abs: https://t.co/d6y4TEFw2h pic.twitter.com/CgMiVVAtyS
CodeRetriever: Unimodal and Bimodal Contrastive Learning for Code Search
— AK (@ak92501) January 27, 2022
abs: https://t.co/lrX5FKqtSc
By fine-tuning with domain/language specified downstream data, CodeRetriever achieves the new sota performance with significant improvement over existing code pre-trained models pic.twitter.com/ZJDAxkljrQ
A Thousand Words Are Worth More Than a Picture:
— AK (@ak92501) January 17, 2022
Natural Language-Centric Outside-Knowledge Visual Question Answering
abs: https://t.co/pmciI9NUSg
TRiG framework outperforms all sota supervised methods by at least 11.1% absolute margin pic.twitter.com/CZkWbL4CPO
PromptBERT: Improving BERT Sentence Embeddings with Prompts
— AK (@ak92501) January 13, 2022
abs: https://t.co/N1EQlKRJAi
github: https://t.co/hrTlbFFks4
Compared to SimCSE, achieve 2.29 and 2.58 points improvements on BERT and RoBERTa respectively under
the unsupervised setting pic.twitter.com/UkhC9F6jTq
I've been working on academic deep learning summarization a bunch. Now we are finally shipping a lot of useful summarization features at https://t.co/2tfrBwhwOc, eg. recipes https://t.co/colxgxPHlv
— Richard Socher (@RichardSocher) December 18, 2021
and
shopping reviews:https://t.co/1OF246JwBI pic.twitter.com/vIU2OVmVAD
We trained a research version of GPT-3 that can search the web, synthesize information, and cite its sources to provide more accurate answers to questions. https://t.co/YEEazt0oLZ
— OpenAI (@OpenAI) December 16, 2021
🚨New paper🚨 SOTA dialogue models are not winning Oscars anytime soon, as they cannot effectively stay in character.
— Jason Weston (@jaseweston) December 14, 2021
We analyze and propose methods to measure & mitigate -- but it's still an open problem.https://t.co/C9SjBPcT4S@shtruk @JackUrbs Arthur Szlam @jaseweston pic.twitter.com/i4V5WkNePT
The three studies explore: Gopher - a SOTA 280B parameter transformer, ethical and social risks, & a new retrieval architecture with better training efficiency.
— DeepMind (@DeepMind) December 8, 2021
1: https://t.co/WDUeFd5DiF
2: https://t.co/cZcWHCg128
3: https://t.co/h9fdMP6C5W (more https://t.co/4QiVDqntTS) 2/
semantic-search-through-wikipedia-with-weaviate - Semantic search through a vectorized Wikipedia (SentenceBERT) with the Weaviate vector search engine https://t.co/eZATb5C9kn
— Python Trending (@pythontrending) December 5, 2021