For my own research and teaching, I follow AI papers. Here is a list of papers I find interesting. I analysed 3 lists (links below) as meta-references for lists of good papers
From that list and also my own interests, here is a set of papers I find interesting in 2020
These influence how AI algorithms could develop in future
1) ADA: Training Generative Adversarial Networks with Limited Data
A method for training a GAN using a very small number of images. Developed by Nvidia.
2) Language Models are Few-Shot Learners
This is the main GPT-3 paper. I have covered before why GPT-3 is disruptive before – Could GPT-3 Change The Way Future AI Models Are Developed and Deplo…
3) Beyond Accuracy: Behavioral Testing of NLP Models with CheckList
Checkist presents a task agnostic way of testing NLP models. Cheklist demonstrates that measures beyond accuracy need to be considered for evaluating some NLP tasks.
4) AutoML-Zero: Evolving Machine Learning Algorithms From Scratch
A paper from Google which aims to show that AutoML can automatically discover complete machine learning algorithms just using basic mathematical operations as building blocks.
5) Towards a Human-like Open-Domain Chatbot
The Meena chatbot from Google can chat about virtually anything in contrast to other chatbots which are more specialized.
1) YOLOv4: Optimal Speed and Accuracy of Object Detection
Improves the YOLO algorithm further in terms of accuracy. AI algorithms continue to improve – even incrementally – but over time all these small research improvements make the technology mainstream. This was a similar trajectory followed by language translation.
2) Unsupervised Translation of Programming Languages
Converts code from a programming language to another without any supervision – for example python functions can be translated to C++ functions.
3) High-Resolution Neural Face Swapping for Visual Effects
This paper from Disney research got a lot of traction. The goal of the paper is to swap the face of a target actor from a source actor while maintaining the actor’s behaviour and performance (for example when ageing the actor).
4) Learning to Cartoonize Using White-box Cartoon Representations
AI can cartoonize any picture or video you feed it in a specified cartoon style.
5) Reconstruct Photorealistic Scenes from Tourists’ Public Photos on the Internet
Using tourists’ public photos from the internet, reconstruct viewpoints of a scene
6) A New Brain-inspired Intelligent System Drives a Car Using Only 19 Control Neurons
AI based on brains of tiny animals (like threadworms – then used to perform complex functions like controlling a self-driving car.
A technique to colorize and restore old black and white images
8) Improving Data‐Driven Global Weather Prediction Using Deep Convolutional Neural Networks on a Cubed Sphere
A significantly-improved data-driven global weather forecasting framework using a deep convolutional neural network (CNN) to forecast several basic atmospheric variables on a global grid.
9) A Distributed Multi-Sensor Machine Learning Approach to Earthquake Early Warning,
A Distributed Multi-Sensor Machine Learning Approach to Earthquake Early Warning based on a new stacking ensemble method
Image source pixabay