ΑΙhub.org
articles
Hot papers on arXiv from the past month – November 2020
Lucy Smith
01 Dec 2020
What’s hot on arXiv? Here are the most tweeted papers that were uploaded onto arXiv during November 2020.
Results are powered by Arxiv Sanity Preserver....
Goodhart’s law, diversity and a series of seemingly unrelated toy problems
BAIR blog
30 Nov 2020
[latexpage]
By Aldo Pacchiano, Jack Parker-Holder, Luke Metz, and Jakob Foerster
Goodhart’s Law is an adage which states the following:
“When a measure becomes a target, it ceases to be a ...
Adapting on the fly to test time distribution shift
BAIR blog
20 Nov 2020
By Marvin Zhang
Imagine that you are building the next generation machine learning model for handwriting transcription. Based on previous iterations of your product, you have identified a key chall...
Efficient graph construction to represent images
Sarath Shekkizhar
12 Nov 2020
Why do we need graphs for image processing?
Image processing over the years has evolved from simple linear averaging filters to highly adaptive non linear filtering operations such as the bilateral f...
On learning language-invariant representations for universal machine translation
ML@CMU
09 Nov 2020
[latexpage]
Figure 1: An encoder-decoder generative model of translation pairs, which helps to circumvent the limitation discussed before. There is a global distribution \(\mathcal{D}\) over the re...
Reinforcement learning is supervised learning on optimized data
BAIR blog
05 Nov 2020
[latexpage]
By Ben Eysenbach and Aviral Kumar and Abhishek Gupta
The two most common perspectives on Reinforcement learning (RL) are optimization and dynamic programming. Methods that compute th...
Hot papers on arXiv from the past month – October 2020
Lucy Smith
02 Nov 2020
What’s hot on arXiv? Here are the most tweeted papers that were uploaded onto arXiv during October 2020.
Results are powered by Arxiv Sanity Preserver....
Improvising with an AI musician
Lucy Smith
28 Oct 2020
For those interested in music and AI, a session on "Human collaboration with an AI musician" at the AI for Good global summit proved to be a real treat. The session included a performance between two ...
Plan2Explore: active model-building for self-supervised visual reinforcement learning
By Oleh Rybkin, Danijar Hafner and Deepak Pathak
[latexpage]
To operate successfully in unstructured open-world environments, autonomous intelligent agents need to solve many different tasks and lea...
AWAC: accelerating online reinforcement learning with offline datasets
BAIR blog
19 Oct 2020
By Ashvin Nair and Abhishek Gupta
[latexpage]
Robots trained with reinforcement learning (RL) have the potential to be used across a huge variety of challenging real world problems. To apply RL...
Generalizing randomized smoothing for pointwise-certified defenses to data poisoning attacks
ML@CMU
16 Oct 2020
We propose a method for making black-box functions provably robust to input manipulations. By training an ensemble of classifiers on randomly flipped training labels, we can use results from randomize...
Five things to know about: making self-driving cars safe
Horizon
12 Oct 2020
By Jonathan O'Callaghan
On 18 September, the European Commission published an independent expert report that looks at some of the outstanding safety and ethical issues around connected and automate...
Hot papers on arXiv from the past month – September 2020
Lucy Smith
07 Oct 2020
What’s hot on arXiv? Here are the most tweeted papers that were uploaded onto arXiv during September 2020.
Results are powered by Arxiv Sanity Preserver....
AI will change the world. Who will change AI? We will.
BAIR blog
05 Oct 2020
By Sophia Stiles
[latexpage]
Editor’s Note: The following blog is a special guest post by a recent graduate of Berkeley BAIR’s AI4ALL summer program for high school students. AI4ALL is a nonp...
Adversarial generation of extreme samples
Lucy Smith
01 Oct 2020
Modelling extreme events in order to evaluate and mitigate their risk is a fundamental goal in many areas, including extreme weather events, financial crashes, and unexpectedly high demand for online ...
ECAI plenary talk: Carme Torras on assistive AI
Lucy Smith
30 Sep 2020
This month saw the European Conference on AI (ECAI 2020) go digital. Included in the programme were five plenary talks. In this article we summarise the talk by Professor Carme Torras who gave an ove...
Exploring exploration: comparing children with RL agents in unified environments
BAIR blog
22 Sep 2020
By Eliza Kosoy, Jasmine Collins and David Chan
Despite recent advances in artificial intelligence (AI) research, human children are still by far the best learners we know of, learning impressive sk...
A round-up of topology-based papers at ICML 2020
Bastian Rieck
17 Sep 2020
With this year’s International Conference on Machine Learning (ICML) being over, it is time to have another instalment of this series. Similar to last year’s post, I shall cover several papers tha...
Can RL from pixels be as efficient as RL from state?
BAIR blog
14 Sep 2020
By Misha Laskin, Aravind Srinivas, Kimin Lee, Adam Stooke, Lerrel Pinto, Pieter Abbeel
[latexpage]
A remarkable characteristic of human intelligence is our ability to learn tasks quickly. Most human...
The role of computer vision in autonomous vehicles
Md Ali Azam
10 Sep 2020
Recent advances in computer vision have revolutionized many areas of research including robotics, automation, and self-driving vehicles. The self-driving car industry has grown markedly in recent year...
High-frequency component helps explain the generalization of convolutional neural networks
ML@CMU
08 Sep 2020
Fig. 1: The central hypothesis: within a dataset with finite samples, there are correlations between the high-frequency component and the “semantic” component of the images. As a result, the model...
Rethinking benchmark systems for machine learning
Common methods applied in the evaluation of model performance share several limitations. A new meta-measure Elo-based Predictive Power (EPP) method addresses these issues....
Hot papers on arXiv from the past month – August 2020
Lucy Smith
02 Sep 2020
What’s hot on arXiv? Here are the most tweeted papers that were uploaded onto arXiv during August 2020.
Results are powered by Arxiv Sanity Preserver....
Decentralized reinforcement learning: global decision-making via local economic transactions
BAIR blog
01 Sep 2020
By Michael Chang and Sidhant Kaushik
[latexpage]
Many neural network architectures that underlie various artificial intelligence systems today bear an interesting similarity to the early computers...
New algorithm follows human intuition to make visual captioning more grounded
Allie McFadden and Machine Learning Center at Georgia Tech
24 Aug 2020
Annotating and labeling datasets for machine learning problems is an expensive and time-consuming process for computer vision and natural language scientists. However, a new deep learning approach is ...
A How-To: reflections on planning virtual science conferences
Juniper Lovato
21 Aug 2020
All Alife 2020 illustrations by Rob Babboni
By Juniper Lovato (general conference chair) and Laurent Hébert-Dufresne (conference co-organizer), Vermont Complex Systems Center, University of Vermont
...
Career advisor systems
Akshay Gugnani and Hemant Misra
18 Aug 2020
Career advisor systems are essentially recommender systems in the space of job searching and career advice. They provide recommendations to candidates with possible career paths and to employers with ...
D4RL: building better benchmarks for offline reinforcement learning
BAIR blog
17 Aug 2020
[latexpage]
By Justin Fu
In the last decade, one of the biggest drivers for success in machine learning has arguably been the rise of high-capacity models such as neural networks along with l...
Maintaining the illusion of reality: transfer in RL by keeping agents in the DARC
ML@CMU
10 Aug 2020
By Benjamin Eysenbach
Reinforcement learning (RL) is often touted as a promising approach for costly and risk-sensitive applications, yet practicing and learning in those domains directly is expens...
#ICML2020 invited talk: Iordanis Kerenidis – “Quantum machine learning : prospects and challenges”
Lucy Smith
07 Aug 2020
The third and final ICML2020 invited talk covered the topic of quantum machine learning (QML) and was given by Iordanis Kerenidis. He took us on a tour of the quantum world, detailing the tools needed...
Simultaneous clustering and representation learning
Ömer Sümer
05 Aug 2020
[latexpage]
The success of deep learning over the last decade, particularly in computer vision, has depended greatly on large training data sets. Even though progress in this area boosted the perfo...
Open compound domain adaptation
BAIR blog
04 Aug 2020
[latexpage]
Figure 1: Domains boundaries are rarely clear. Therefore, it is hard to set up definite domain descriptions for all possible domains.
By Zhongqi Miao and Ziwei Liu
The World i...
Hot papers on arXiv from the past month – July 2020
Lucy Smith
03 Aug 2020
What’s hot on arXiv? Here are the most tweeted papers that were uploaded onto arXiv during July 2020.
Results are powered by Arxiv Sanity Preserver....
#ICML2020 invited talk: Brenna Argall – “Human and machine learning for assistive autonomy”
Lucy Smith
31 Jul 2020
The second invited talk at ICML2020 was given by Brenna Argall. Her presentation covered the use of machine learning within the domain of assistive machines for rehabilitation. She described the effor...
#ICML2020 invited talk: Lester Mackey – “Doing some good with machine learning”
Lucy Smith
28 Jul 2020
There were three invited talks at this year's virtual ICML. The first was given by Lester Mackey, and he highlighted some of his efforts to do some good with machine learning. During the talk he also ...
In defense of weight-sharing for neural architecture search: an optimization perspective
ML@CMU
27 Jul 2020
[latexpage]
Figure 1: How DARTS and other weight-sharing methods replace the discrete assignment of one of four operations \(o\in O\) to an edge \(e\) with a \(\theta\)-weighted combination of their ...
