Every two weeks, we find the most interesting data science links from around the web and collect them in Data Science Briefings, the DataMiningApps newsletter. Subscribe now for free if you want to be the first to get up to speed on interesting resources.
- Monitoring Machine Learning Models in Production
Once you have deployed your machine learning model to production it rapidly becomes apparent that the work is not over. - Why We Need DevOps for ML Data
Getting machine learning (ML) into production is hard. In fact, it’s possibly an order of magnitude harder than getting traditional software deployed. - Data Drift Detection for Image Classifiers
“In the context of machine learning, we consider data drift to be the change in model input data that leads to a degradation of model performance.” - Can robots help us overcome the coronavirus health crisis and lockdown?
“The widespread use of robots has long been debated, in part because many critics see them as a threat to jobs and livelihoods. In time of crisis and catastrophes, however, robots and AI have proved that they can provide crucial assistance.” - Open Source Virtual Background
“Since I do not have a green screen I decided to simply implement my own background removal, which was obviously better than cleaning my apartment or just using my laptop all the time” - Facebook AI, AWS partner to release new PyTorch libraries
Facebook AI and AWS engineers have partnered to develop new libraries targeted at large-scale elastic and fault-tolerant model training and high-performance PyTorch model deployment. - This open-source program lets you run deepfakes on live video calls
With some programming knowledge and solid hardware, you can use Avatarify to fake your Zoom calls… or someone else’s. - A Graph Convolutional Neural Network Approach to Antibiotic Discovery
“In an age when bacterial infections are developing resistance to common antibiotics, the discovery of a new and potentially powerful antibiotic is news in itself. But what makes a recent breakthrough truly revolutionary is that the promising molecule—called halicin—was discovered using deep learning.” - Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild
“We propose a method to learn weakly symmetric deformable 3D object categories from raw single-view images, without ground-truth 3D, multiple views, 2D/3D keypoints, prior shape models or any other supervision.” - Specification gaming: the flip side of AI ingenuity
Specification gaming is a behaviour that satisfies the literal specification of an objective without achieving the intended outcome. - How to know if artificial intelligence is about to destroy civilization
These canaries in the coal mines of AI would be signs that superintelligent robot overlords are approaching - OpenAI Jukebox
We’re introducing Jukebox, a neural net that generates music, including rudimentary singing, as raw audio in a variety of genres and artist styles. We’re releasing the model weights and code, along with a tool to explore the generated samples. - Generating MIDI with 10k–30k Context Windows
Successful pieces are possible, and are musically far more diverse than the folk ABC corpus, with many pleasingly complex samples. - More Descent, Less Gradient
Your regression may only need one gradient step. Really. - AI Being Incorporated in Architecture, Is Potentially Transforming
AI is beginning to transform architecture as its expanding uses include evaluating the efficiency of construction and simulating movement in a space. - micrograd
A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API - Quake 1 Movement Physics Reinforcement Learning
“I developed a facsimile of Quake player movement physics in Python, wrapped with an OpenAI gym.Env. I then trained an agent for this using RLLib’s implementation of PPO. In the end the agent was able to beat the current human speedrun record on a speed running practice map.” - Selfie 2 Waifu
Based on U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation - Estimating COVID-19’s Rt in Real-Time
A modified version of a solution created by Bettencourt & Ribeiro 2008 to estimate real-time Rt using a Bayesian approach. - Coronavirus: How AI, 3D printing and blockchain can help overcome supply problems in a crisis
“Investing in AI, 3D printing and blockchain can help businesses be more responsive to changes in supply and demand and make their supply chains more resilient.” - How to run R 4.0 in Docker — and 3 cool new R 4.0 features
See how to test drive R 4.0 in a Docker container, plus a look at three new R 4.0 features for color palettes and strings - Barclays Innovating in use of AI in Banking
“Barclays Bank is emerging as an innovator in the use of AI in financial services. The UK bank, ranked 20th on the S&P Global’s list of the top 100 banks, works with suppliers of AI products and services more than it develops AI applications in house”