AI, Machine Learning and Deep Learning News

Categories Machine Learning

This page serves as a chronological catalogue used when discussing Machine and Deep Learning news with the Code Academy class in 2020 – 2021.


  1. Is there a ML / DL model to detect logical fallacies, some research: and


  1. Hundreds of AI tools have been built to catch covid. None of them helped: There are some curious cases of GIGO mentioned, like: bias introduced by pictorial artifacts like timestamps on the picture being added by some hospital in a certain place on the picture, data with covid positive patients being intermixed with data from children, dataset where standing people’s scans are distributed with patient’s scanned while laying down.
  2. >95% accuracy in Covid prediction: TODO – need to find the dataset and try to see what kind of results could we get in class.
  3. Lithuanian companies working with ML/DL:


  1. CPython will get faster:


  1. Awesome python repository of awesome python libz and frameworks:
  2. Facebook AI: “High-performance speech recognition with no supervision at all”:
  3. News concentrator:


  1. Summary of Google I/O AI improvements:


  1. 2 big inventions in DL 2020:
  2. Faceception:
  3. iDenfy and Manu Analytics:
  4. Mercedes AI:


  1. Silent ML


  1. Mac mini teardown:


  1. Is Macbook M1 good for software engineering:
  2. Great for datascience in the comments on this video:
  3. Bench:


  1. Real time person remover:
  2. Yet Another EfficientDet Pytorch:
  3. Facebook DETR:
  4. YOLOv5:


  1. Some data analysis techniques applies for election fraud investigation:
  2. Benford’s law applied to corona:
  3. More, in human:


  1. Interesting traffic monitoring:
  2. ML flashcards: (proposed here not as a advertisement, but as a pedagogical idea)


  1. Only when organizations add the ability to learn with AI do significant benefits become likely
  2. Neural Filters in Photoshop


  1. An interesting question to track over the years in the autonomous self driving cars is who is liable. Tesla has several features in it’s autopilot: lane-keep, navigate on autopilot (highway autonomy), smart summon, active safety, pedestrian aeb and full self-driving (only available on development builds). In the cases of navigate and smart summon – who is responsible for a “fender-bender”? This article answers that:


  1. Transformers replacing CNNs for image recognition
  2. Learning with tiny amounts of data with LOS

Leave a Reply