Machine Learning Research
Moderating the ML Roller Coaster: A technique to avoid double descent in AI
Wait a minute — we added training data, and our model’s performance got worse?! New research offers a way to avoid so-called double descent.
Machine Learning Research
Machine Learning Research
Preserving Detail in Image Inputs: Better image compression for computer vision datasets
Given real-world constraints on memory and processing time, images are often downsampled before they’re fed into a neural network. But the process removes fine details, and that degrades accuracy. A new technique squeezes images with less compromise.
Machine Learning Research
Outside the Norm: Batch normalization contributes to neural network accuracy.
Batch normalization is a technique that normalizes layer outputs to accelerate neural network training. But new research shows that it has other effects that may be more important.
Machine Learning Research
Beyond Neural Architecture Search: AutoML-Zero is a meta-algorithm for classification.
Faced with a classification task, an important step is to browse the catalog of machine learning architectures to find a good performer. Researchers are exploring ways to do it automatically.
Machine Learning Research
Seeing the See-Through: ClearGrasp allows robots to grab see-through objects.
Glass bottles and crystal bowls bend light in strange ways. Image processing networks often struggle to separate the boundaries of transparent objects from the background that shows through them. A new method sees such items more accurately.
Machine Learning Research
Optimize Your Training Parameters: Research on finding a neural net's optimal batch size
Last week we reported on a formula to determine model width and dataset size for optimal performance. A new paper contributes equations that optimize some training parameters.
Machine Learning Research
Rightsizing Neural Nets: An equation for predicting optimal data and model size
How much data do we want? More! How large should the model be? Bigger! How much more and how much bigger? New research estimates the impact of dataset and model sizes on neural network performance.
Machine Learning Research
Deep Learning Finds New Antibiotic: Researchers used AI to identify a promising new antibiotic.
Chemists typically develop new antibiotics by testing close chemical relatives of tried-and-true compounds like penicillin. That approach becomes less effective, though, as dangerous bacteria evolve resistance to those very chemical structures. Instead, researchers enlisted neural networks.
Machine Learning Research
X Marks the Dataset: Radioacive data helps trace a model's training corpus.
Which dataset was used to train a given model? A new method makes it possible to see traces of the training corpus in a model’s output.
Machine Learning Research
Self-Supervised Simplicity: Image classification with simple contrastive learning (SimCLR)
A simple linear classifier paired with a self-supervised feature extractor outperformed a supervised deep learning model on ImageNet, according to new research.
Machine Learning Research
Guest Speaker: Deepfake method syncs up mouth movements with words.
Deepfake videos in which one person appears to speak another’s words have appeared in entertainment, advertising, and politics. New research ups the ante for an application that enables new forms of both creative expression and
Machine Learning Research
Less Labels, More Learning: Improved small data performance with combined techniques
In small data settings where labels are scarce, semi-supervised learning can train models by using a small number of labeled examples and a larger set of unlabeled examples. A new method outperforms earlier techniques.