Publicaciones

As AI improves at a rapid pace, we’re able to detect more hate speech, bullying, and harassing content than ever before. Our engineers worked to improve the way our AI systems analyze comments, considering both the comments and their context,
providing a better and deeper understanding of language as well as the ability to combine analyses of images, text, and other details contained in a post. Learn more: https://ai.facebook.com/…/ai-gets-better-every-day-heres-…/…

A medida que la IA mejora a un ritmo rápido, podemos detectar más lenguaje de odio, intimidación y contenido hostigador que nunca antes. Nuestros ingenieros trabajaron para mejorar la forma en que nuestros sistemas de IA analizan comentarios, considerando tanto los comentarios como su contexto, proporcionar una mejor y más profunda comprensión del lenguaje, así como la capacidad de combinar análisis de imágenes, texto y otros detalles contenidos en una publicación. Más información: https://ai.facebook.com/blog/ai-gets-better-every-day-heres-what-that-means-for-stopping-hate-speech/?utm_source=hootsuite&utm_medium=&utm_term=&utm_content=&utm_campaign=
Traducido
We’re working to improve our AI systems that identify and analyze hate speech in comments. Millions of pieces of content were detected in 2020 and removed for violating our policies before anyone reported them, detecting more hate speech, bullying, and harassment content than ever before.
ai.facebook.com
AI gets better every day. Here’s what that means for stopping hate speech
We’re working to improve our AI systems that identify and analyze hate speech in comments. Millions of pieces of content were detected in 2020 and removed for violating our policies before anyone reported them, detecting more hate speech, bullying, and harassment content than ever before.
We’re working to improve our AI systems that identify and analyze hate speech in comments. Millions of pieces of content were detected in 2020 and removed for violating our policies before anyone reported them, detecting more hate speech, bullying, and harassment content than ever before.

Introducing TT-Rec, a new method to dramatically shrink the size of cutting-edge Deep Learning Recommendation Models, which are very memory intensive. TT-Rec can also make it much easier to deploy these models at scale. We look forward to presenting our research at MLSys Conference 2021 and are sharing the code now so others can improve their DLRMs. https://arxiv.org/abs/2101.11714

Presentamos TT-Rec, un nuevo método para reducir dramáticamente el tamaño de modelos de recomendación de aprendizaje profundo de vanguardia, que son muy intensivos de memoria. TT-Rec también puede facilitar el despliegue de estos modelos a escala. Esperamos presentar nuestra investigación en MLSys Conference 2021 y estamos compartiendo el código ahora para que otros puedan mejorar sus DLRMs. https://arxiv.org/abs/2101.11714
Traducido
Puede ser una imagen de texto que dice "14 GB 12 12 Requirement の 10 0 Memory 2 TT-Rec Compression Ratio 4x 48.2x 117.3x TT-Rec Compression Ratio 2.5x 21.8x 95.5x Baseline 3TT- Emb. 5TT- Emb. 7n- Baseline 3 TT- Emb. Emb. Kaggle 5TT- Emb. 7T- Emb. Terabyte"
Ver todas
Vídeos
Model-Based Inverse Reinforcement Learning from Visual Demonstrations
151
11
We’re releasing fairmotion, a library to help AI researchers use motion capture data for computer graphics and robotics. It provides tools to load, process and visualize motion, and demonstrates its utility with example tasks. We’ve used the library in our work presented at SIGGRAPH 2020 on controlling diverse behaviors for physically simulated characters. Get the code and paper: http://github.com/facebookresearch/fairmotion https://research.fb.com/publications/a-scalable-approach-to-control-diverse-behaviors-for-physically-simulated-characters/
351
16
AI_Integrity_Rio_animation_V2.mp4
169
9
Ver todos
Fotos
Puede ser una imagen de texto que dice "14 GB 12 12 Requirement の 10 0 Memory 2 TT-Rec Compression Ratio 4x 48.2x 117.3x TT-Rec Compression Ratio 2.5x 21.8x 95.5x Baseline 3TT- Emb. 5TT- Emb. 7n- Baseline 3 TT- Emb. Emb. Kaggle 5TT- Emb. 7T- Emb. Terabyte"
Puede ser una imagen de alimentos y texto que dice "Research in Brief FACEBOOK AI"
Puede ser una imagen de rayos X
Ver todas