A summary of recent updates in the field of Artificial Intelligence (AI) for week 45 of 2020. For more regular updates and related discussions, join us in the Artificial Intelligence Community. As always, thanks for your support for sharing it with your friends and colleagues on social media.

Industry updates for week 45 of 2020

This section is for recent industry updates related to Artificial Intelligence.

  • German startup gets 3.5k stars: The German startup, AMAI developed a road map for AI and received an overwhelming response with 3,500 stars on Github. The roadmap is user friendly and customizable. Here is full coverage on the same.
  • Interview of Geoff Hinton: Hinton’s views on how AI has been leading every market application and his long-term belief in Deep learning have been on every researcher’s mind now. While speaking at MIT’s emTech conference, Hinton delivered his thoughts on data scaling, GPT-3, and of course future of AI. Here is the entire conversation for a better read. 
  • Super héroes of Deep Learning: If you are a comic fan and a nerd who is also obsessed with AI and Machine learning, here is your personalized e-copy of volume 1 of superheroes of Deep Learning. 
  • Did Trump Tweet this?Here is an application for Android users who have a keen interest in evaluating a tweet for real and fake ones by Donald Trump. No harm in trying it out and sure some fun is guaranteed.
  • Google, OpenAI, and DeepMind use “behavior priors”: To boost the performance of RL agents, the importance of prior knowledge will enhance their abilities to solve complex problems. So behavior priors help to capture movement and interaction patterns to speedup complex tasks and converge effectively. To know more, visit here.
  • MIT, Harvard, UChicago, and Diffeo use Bayesian Interference: The Bayesian delegation solves the basic challenge faced by any multi-agent learning mechanism that helps them coordinate on the fly. Here is the entire post.

Articles for week 45 of 2020

This section is for recent Blog Posts/Tutorials related to Artificial Intelligence.

  • Problems of Bias: The recent misclassification of arresting a black man as a criminal implies, the biased dataset problem. This issue needs to be addressed and this post serves the same purpose. To read how to deal with bias, visit here.
  • Consumer Segmentation: Quilt.AI collected data from parents’ data available on social media and forums to analyze keywords and pose common concerns in the form of queries. Here are entire data statistics from what they learned from such a study.
  • AI Emotion Recognition: The article covers a comprehensive study of generic machine learning-based emotion recognition and better ways to do so. To know more, visit here.
  • Fast RCNN for object detection: The article covers a comparison of fast RCNN with state-of-the-art SPPNet and RCNN. To read the full post visit here.
  • Batch Normalization: It is a detailed article on Batch Normalization with the description in three styles according to the reader i.e. 30-second read, 3minute read, or a comprehensive one. It covers key elements with basic implementation in Pytorch.
  • Juke Box: OpenAI’s Jukebox is an AI music generation platform. The model is developed using Piano Transformer from Tensorflow. Here is the entire post.
  • Roboat: The researchers at MIT developed an autonomous vehicle to travel across the water body called the “Roboat”. It can carry up to 2 people and navigate efficiently. Here is the detailed news.
  • AI for mental health: The diagnosis and treatment of mental health is now possible through AI. Though this does not hamper the income of health professionals it surely helps one keep a check as most of us are shy consulting for our mental/emotional problems. Here is a special post mentioning applications dedicated to mental health diagnosis.
  • Facebook’s misinformation detector: Months before the election, Facebook developed a system that detected fake accounts and posts. Though there is no evidence they used this feature but they certainly showed promising results. Here is the news post on the same.
  • Twitter sentiment analysis: Building a twitter-based sentiment analysis in easy steps is possible now. To give it a try, visit here.
  • Iris’s landmark tracking: The Iris tracking being difficult due to the physical appearance of the eye and its surrounding obstacles such as eyeglasses and blinking of an eye, researchers have developed Iris trackers using Mediapipe and Tensorflow.js. To know more, visit here.
  • Learning Shortlist for Object detection: This post provides all the shortlisted resources required to start learning object detection in computer vision. 
  • Face Recognition face-off: Oh ho, cops be careful before you mistakenly ill-treat any civilian because they are using face recognition to chase you down. Here is the Batch’s featuring this week’s line opener for this week.


This section is for recent Tools/Applications related to Artificial Intelligence.


This section is for recent Lectures/Conferences/Webinars related to Artificial Intelligence.

  • Deep Learning Introduction: It is a video lecture series by Andreas Maier from FAU.


This section is for recent research published related to Artificial Intelligence.

  • MEDMNIST classification: The researchers at SJTU have developed MedMNIST a dataset of 10 pre-collected datasets for carrying out research on Medical Imaging. To read news coverage visit here and the research article is available here while the Github repository is also accessible.
  • Data Augmentation via Structured Adversarial Perturbations: The Adversarial Perturbations is an alternative for data augmentation using photometric and geometric images. To read the entire article read here.


This section is for updates for other various related topics to Artificial Intelligence.

  • Cylab’s Intrusion Detection Systems: MIT’s Cylab has developed the fastest open-source intrusion detection system named Pigasus. The system can achieve 100Gbps using a single address. Here are a news article and Github link. 
  • Resources for GANs and deep fakes: If you are searching for all the resources in one place, here is the Github link for reference on GANs and deep fakes.
  • Original Transformer model: The Github repository for building an Original Transformer model using Pytorch can be accessed from here.
  • Papers are all you need: To download your choice of research articles for a specific area with an author in my, this Github repo just does that for you. To mind you, it is currently working for NeurIPS, and in case anyone wishes to extend it to other conferences and journals, he/she may contact the repository owner.

That’s all for the week 45 of 2020. In case you think we have missed out something and that is significant to add, please free to reach out and submit to us via our Contact form or via Social media channels: Twitter and Facebook.