DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification
In this research, Yongming Rao et al. propose a dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input. A lightweight prediction module can estimate the importance score of each token given the current features. The module is added to different layers to prune redundant tokens hierarchically.
Web Page — https://bit.ly/3dgESlq
Paper — https://bit.ly/3xWf5XJ
Code — https://bit.ly/3jlUtUK
In this research, Yongming Rao et al. propose a dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input. A lightweight prediction module can estimate the importance score of each token given the current features. The module is added to different layers to prune redundant tokens hierarchically.
Web Page — https://bit.ly/3dgESlq
Paper — https://bit.ly/3xWf5XJ
Code — https://bit.ly/3jlUtUK
dynamicvit.ivg-research.xyz
DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification
Yongming Rao1 Wenliang Zhao1 Benlin Liu2
Jiwen Lu1 Jie Zhou1 Cho-Jui Hsieh2
1Tsinghua University 2University of California, Los Angeles
[Paper (arXiv)] …
Yongming Rao1 Wenliang Zhao1 Benlin Liu2
Jiwen Lu1 Jie Zhou1 Cho-Jui Hsieh2
1Tsinghua University 2University of California, Los Angeles
[Paper (arXiv)] …
Consistent Instance False Positive Improves Fairness in Face Recognition
In this paper, Xingkun Xu et al. propose a false positive rate penalty loss, a novel method to mitigate face recognition bias by increasing the consistency of instance False Positive Rate (FPR). The method requires no demographic annotations, allowing to mitigate bias among demographic groups divided by various attributes.
Paper — https://bit.ly/361fHiQ
Code — https://bit.ly/2UKAqoF
In this paper, Xingkun Xu et al. propose a false positive rate penalty loss, a novel method to mitigate face recognition bias by increasing the consistency of instance False Positive Rate (FPR). The method requires no demographic annotations, allowing to mitigate bias among demographic groups divided by various attributes.
Paper — https://bit.ly/361fHiQ
Code — https://bit.ly/2UKAqoF
The FLORES-101 Data Set: Helping Build Better Translation Systems Around the World
Building on the success of machine translation systems like M2M-100, Facebook AI has open-sourced FLORES-101, a many-to-many evaluation data set covering 101 languages from all over the world, to enable researchers to rapidly test and improve upon multilingual translation models like M2M-100. In this article, you’ll delve into its basics.
https://bit.ly/3hja98z
Building on the success of machine translation systems like M2M-100, Facebook AI has open-sourced FLORES-101, a many-to-many evaluation data set covering 101 languages from all over the world, to enable researchers to rapidly test and improve upon multilingual translation models like M2M-100. In this article, you’ll delve into its basics.
https://bit.ly/3hja98z
Facebook
The FLORES-101 data set: Helping build better translation systems around the world
Today we are open-sourcing FLORES-101, a first-of-its-kind, many-to-many evaluation data set covering 101 languages (10,100 translation directions) from all over the world.
Multivariate Probabilistic Regression with Natural Gradient Boosting
Natural Gradient Boosting (NGBoost) is a new method proposed by the researchers. It is based on nonparametrically modeling the conditional parameters of the multivariate predictive distribution. The method is robust, works out-of-the-box without extensive tuning, is modular with respect to the assumed target distribution, and performs competitively in comparison to existing approaches.
Paper — https://bit.ly/3haEF5C
Code — https://bit.ly/3qC1SB3
Natural Gradient Boosting (NGBoost) is a new method proposed by the researchers. It is based on nonparametrically modeling the conditional parameters of the multivariate predictive distribution. The method is robust, works out-of-the-box without extensive tuning, is modular with respect to the assumed target distribution, and performs competitively in comparison to existing approaches.
Paper — https://bit.ly/3haEF5C
Code — https://bit.ly/3qC1SB3
Data Phoenix Rises
We at Data Science Digest have always strived to ignite the fire of knowledge in the AI community. We’re proud to have helped thousands of people to learn something new and give you the tools to push ahead. And we’ve not been standing still, either.
Please meet Data Phoenix, a Data Science Digest rebranded and risen anew from our own flame. Our mission is to help everyone interested in Data Science and AI/ML to expand the frontiers of knowledge. More news, more updates, and webinars (!) are coming. Stay tuned!
Data Phoenix Digest — 01.07.2021
The new issue of Data Phoenix Digest is here! AI that helps write code, EU’s ban on biometric surveillance, genetic algorithms for NLP, multivariate probabilistic regression with NGBoosting, alias-free GAN, MLOps toys, and more…
https://bit.ly/3h722gE
Join 👉@DataPhoenix
We at Data Science Digest have always strived to ignite the fire of knowledge in the AI community. We’re proud to have helped thousands of people to learn something new and give you the tools to push ahead. And we’ve not been standing still, either.
Please meet Data Phoenix, a Data Science Digest rebranded and risen anew from our own flame. Our mission is to help everyone interested in Data Science and AI/ML to expand the frontiers of knowledge. More news, more updates, and webinars (!) are coming. Stay tuned!
Data Phoenix Digest — 01.07.2021
The new issue of Data Phoenix Digest is here! AI that helps write code, EU’s ban on biometric surveillance, genetic algorithms for NLP, multivariate probabilistic regression with NGBoosting, alias-free GAN, MLOps toys, and more…
https://bit.ly/3h722gE
Join 👉@DataPhoenix
What Is MLOps? — Everything You Must Know to Get Started
MLOps is a buzzword right now. Everyone talks about it; everybody wants to implement it and drive MLOps transformations. If you’re interested in what MLOps is too, this article will provide a scoop of ML systems development lifecycle and explain why you need MLOps.
https://bit.ly/3dB9gau
MLOps is a buzzword right now. Everyone talks about it; everybody wants to implement it and drive MLOps transformations. If you’re interested in what MLOps is too, this article will provide a scoop of ML systems development lifecycle and explain why you need MLOps.
https://bit.ly/3dB9gau
Medium
What is MLOps — Everything You Must Know to Get Started
A complete walkthrough of ML systems development lifecycle and need for MLOps
To Retrain, or Not to Retrain? Let’s Get Analytical About ML Model Updates
In this ML 101 article, you’ll find answers to questions like, «How often should I retrain a model?», «Should I retrain the model now?», and «Should I retrain, or should I update the model?». Dig in for an easy but important piece to read!
https://bit.ly/3qVwDRL
In this ML 101 article, you’ll find answers to questions like, «How often should I retrain a model?», «Should I retrain the model now?», and «Should I retrain, or should I update the model?». Dig in for an easy but important piece to read!
https://bit.ly/3qVwDRL
Evidentlyai
To retrain, or not to retrain? Let’s get analytical about ML model updates.
Is it time to retrain your machine learning model? Even though data science is all about… data, the answer to this question is surprisingly often based on a gut feeling. Can we do better?
PlanSys2: A Planning System Framework for ROS2
In this paper, the researchers reveal the ROS2 Planning System (PlanSys2), a framework for symbolic planning that incorporates novel approaches for execution on robots working in demanding environments. PlanSys2 aims to be the reference task planning framework in ROS2, the latest version of the {\em de facto} standard in robotics software development.
Paper — https://bit.ly/3qRr8U4
Code
— https://bit.ly/3yy3Sx5
In this paper, the researchers reveal the ROS2 Planning System (PlanSys2), a framework for symbolic planning that incorporates novel approaches for execution on robots working in demanding environments. PlanSys2 aims to be the reference task planning framework in ROS2, the latest version of the {\em de facto} standard in robotics software development.
Paper — https://bit.ly/3qRr8U4
Code
— https://bit.ly/3yy3Sx5
A Discourse on Reinforcement Learning [Part 1]
This is the first of the 3-article series «A Discourse on Reinforcement Learning» that kicks off with a holistic overview of Reinforcement Learning with an expansive setting. Save the article not to miss parts 2 and 3 about more advanced RL topics.
https://bit.ly/3ApUZr0
This is the first of the 3-article series «A Discourse on Reinforcement Learning» that kicks off with a holistic overview of Reinforcement Learning with an expansive setting. Save the article not to miss parts 2 and 3 about more advanced RL topics.
https://bit.ly/3ApUZr0
Medium
A Discourse on Reinforcement Learning
Part I — AN EXPANSIVE SETTING
The MultiBERTs: BERT Reproductions for Robustness Analysis
In this paper, the international team of researchers introduce MultiBERTs: a set of 25 BERT-base checkpoints, trained with similar hyper-parameters as the original BERT model but differing in random initialization and data shuffling. The aim is to enable researchers to draw robust and statistically justified conclusions about pre-training procedures.
Paper — https://bit.ly/3ylSVyh
Code — https://bit.ly/3dGW3wY
In this paper, the international team of researchers introduce MultiBERTs: a set of 25 BERT-base checkpoints, trained with similar hyper-parameters as the original BERT model but differing in random initialization and data shuffling. The aim is to enable researchers to draw robust and statistically justified conclusions about pre-training procedures.
Paper — https://bit.ly/3ylSVyh
Code — https://bit.ly/3dGW3wY
An Introduction to Object Detection with Deep Learning
In this article, you’ll learn the basics of using object-detection deep learning networks. It features a review of CNNs, object detection datasets, R-CNN models, and YOLO. If you’re looking for a review of deep learning architectures for object detection, this one’s for you.
https://bit.ly/3yppLOY
In this article, you’ll learn the basics of using object-detection deep learning networks. It features a review of CNNs, object detection datasets, R-CNN models, and YOLO. If you’re looking for a review of deep learning architectures for object detection, this one’s for you.
https://bit.ly/3yppLOY
TechTalks
An introduction to object detection with deep learning
Learn how deep neural networks detect objects in images in this primer on object detection with deep learning.
AutoFormer: Searching Transformers for Visual Recognition
AutoFormer is a new one-shot architecture search framework for vision transformer search. It entangles the weights of different blocks in the same layers during supernet training. The trained supernet allows thousands of subnets to be well-trained. Their performance with weights inherited from the supernet is comparable to those retrained from scratch.
Paper — https://bit.ly/3dKRiCl
Code — https://bit.ly/3As6zln
AutoFormer is a new one-shot architecture search framework for vision transformer search. It entangles the weights of different blocks in the same layers during supernet training. The trained supernet allows thousands of subnets to be well-trained. Their performance with weights inherited from the supernet is comparable to those retrained from scratch.
Paper — https://bit.ly/3dKRiCl
Code — https://bit.ly/3As6zln
Vectorization Techniques in NLP [Guide]
This comprehensive guide by the Neptune team will help you dig deep into Natural Language Processing and explore all the main branches of word embeddings, starting from naive count-based methods to sub-word level contextual embeddings.
https://bit.ly/3hBiODH
This comprehensive guide by the Neptune team will help you dig deep into Natural Language Processing and explore all the main branches of word embeddings, starting from naive count-based methods to sub-word level contextual embeddings.
https://bit.ly/3hBiODH
neptune.ai
Vectorization Techniques in NLP [Guide] - neptune.ai
Natural Language is how we, humans, exchange ideas and opinions. There are two main mediums for natural language – speech and text. Listening and reading are effortless for a healthy human, but they’re difficult for a machine learning algorithm. That’s why…
DivergentNets: Medical Image Segmentation by Network Ensemble
In this paper, the team of researchers explore new methods of detection of colon polyps with machine learning. They propose DivergentNets, an ensemble of such well-known segmentation models as UNet++, FPN, DeepLabv3, and DeepLabv3+, to produce more generalizable medical image segmentation masks.
Paper — https://bit.ly/3jMNG6M
Code — https://bit.ly/3xqB21i
In this paper, the team of researchers explore new methods of detection of colon polyps with machine learning. They propose DivergentNets, an ensemble of such well-known segmentation models as UNet++, FPN, DeepLabv3, and DeepLabv3+, to produce more generalizable medical image segmentation masks.
Paper — https://bit.ly/3jMNG6M
Code — https://bit.ly/3xqB21i
Continuously Improving Recommender Systems for Competitive Advantage Using NVIDIA Merlin and MLOps
In this article, you’ll explore how to use NVIDIA Merlin, an application framework that accelerates all phases of recommender system development on NVIDIA GPUs, to implement a complete MLOps pipeline.
https://bit.ly/2Uonwwy
In this article, you’ll explore how to use NVIDIA Merlin, an application framework that accelerates all phases of recommender system development on NVIDIA GPUs, to implement a complete MLOps pipeline.
https://bit.ly/2Uonwwy
NVIDIA Developer Blog
Continuously Improving Recommender Systems for Competitive Advantage Using NVIDIA Merlin and MLOps | NVIDIA Developer Blog
This posts shares how NVIDIA Merlin components fit into a complete MLOps pipeline to operationalize a recommendation system, and continuously deliver improvements in production.
Are you subscribed to our weekly newsletter yet? What are you waiting for, then? Fill in your email and get instant access to all the AI/ML goodies in one go. Looking forward to having you as one of our amazing subscribers!
https://dataphoenix.info/tag/digest/
https://dataphoenix.info/tag/digest/
Last week, we rebranded and restarted our digest. What does it mean, exactly?
For you, it means that you’re going to get updated about all things happening in the AI world on a new level. We’re hyped and ready to give you more and be there for you when you need guidance. Our weekly newsletter will be sent out as usual every Thursday (so no worries), but we’d also love to coop more with the community. We’ve planned tons of new activities for you, so make sure that you follow us on social media — and be ready to get energized!
https://bit.ly/2VjWN4O
For you, it means that you’re going to get updated about all things happening in the AI world on a new level. We’re hyped and ready to give you more and be there for you when you need guidance. Our weekly newsletter will be sent out as usual every Thursday (so no worries), but we’d also love to coop more with the community. We’ve planned tons of new activities for you, so make sure that you follow us on social media — and be ready to get energized!
https://bit.ly/2VjWN4O