[KongchangAI]
CompanyFAIR / Facebook AI Research

Facebook AI

The artificial intelligence research organization under Meta (formerly Facebook), focused on frontier AI research including deep learning and natural language processing. It conducts and publishes work spanning model architectures, representation learning, computer vision and language understanding, and has released widely used open-source frameworks and pretrained models. The organization has operated under the name FAIR (Facebook AI Research) and is referred to as Meta AI Research following the company's rebranding.

Timeline (last 90 days)

Nov 16

In Facebook AI's "negotiation agents" experiment, agents were observed abandoning natural language in favor of self-invented encodings.

Unverified50%
Nov 10

XLM-RoBERTa is a multilingual pre-trained language model released by Facebook AI, with the large version having about 560 million parameters; BGE-M3 is a multilingual embedding model released by the Beijing Academy of Artificial Intelligence with about 570 million parameters; multilingual-e5-large-instruct is a multilingual instruction-tuned version of Microsoft's E5 series with about 560 million parameters.

Unverified50%
Nov 6

XLM-RoBERTa-large is a multilingual pretrained encoder released by Facebook AI (now Meta AI), based on the RoBERTa architecture, trained on a CommonCrawl corpus covering 100 languages, with about 560 million parameters.

Unverified50%
Nov 3

RoBERTa is a pretrained language model proposed by Facebook AI in 2019 and can be seen as an optimized version of BERT, removing the next-sentence prediction task, using larger batches and more data, and adopting a dynamic masking mechanism.

Unverified50%
Oct 23

Representative early work on social agents includes Facebook AI's PersonaChat dataset and the BlenderBot series.

Unverified50%
Sep 18

RoBERTa is a language model trained by Facebook AI with optimizations built on top of BERT.

Unverified50%
Sep 12

BART was introduced by Facebook AI in 2019 and was designed for text denoising and generation tasks, with the encoder responsible for understanding input and the decoder responsible for generating output.

Unverified50%
Sep 4

Facebook AI introduced DETR (DEtection TRansformer) in 2020.

Unverified50%

All Facts (8)

Unverified

In Facebook AI's "negotiation agents" experiment, agents were observed abandoning natural language in favor of self-invented encodings.

50%
Unverified

XLM-RoBERTa is a multilingual pre-trained language model released by Facebook AI, with the large version having about 560 million parameters; BGE-M3 is a multilingual embedding model released by the Beijing Academy of Artificial Intelligence with about 570 million parameters; multilingual-e5-large-instruct is a multilingual instruction-tuned version of Microsoft's E5 series with about 560 million parameters.

50%
Unverified

XLM-RoBERTa-large is a multilingual pretrained encoder released by Facebook AI (now Meta AI), based on the RoBERTa architecture, trained on a CommonCrawl corpus covering 100 languages, with about 560 million parameters.

50%
Unverified

RoBERTa is a pretrained language model proposed by Facebook AI in 2019 and can be seen as an optimized version of BERT, removing the next-sentence prediction task, using larger batches and more data, and adopting a dynamic masking mechanism.

50%
Unverified

Representative early work on social agents includes Facebook AI's PersonaChat dataset and the BlenderBot series.

50%
Unverified

RoBERTa is a language model trained by Facebook AI with optimizations built on top of BERT.

50%
Unverified

BART was introduced by Facebook AI in 2019 and was designed for text denoising and generation tasks, with the encoder responsible for understanding input and the decoder responsible for generating output.

50%
Unverified

Facebook AI introduced DETR (DEtection TRansformer) in 2020.

50%

Source Articles