AI/ML Techniques

Why Initialize a Neural Network with Random Weights?

Tweet Tweet Share Share Last Updated on August 15, 2022 The weights of artificial neural networks must be initialized to…

2 years ago

When to Use MLP, CNN, and RNN Neural Networks

Tweet Tweet Share Share Last Updated on August 15, 2022 What neural network is appropriate for your predictive modeling problem?…

2 years ago

Difference Between a Batch and an Epoch in a Neural Network

Tweet Tweet Share Share Last Updated on August 15, 2022 Stochastic gradient descent is a learning algorithm that has a…

2 years ago

Using Depthwise Separable Convolutions in Tensorflow

Tweet Tweet Share Share Last Updated on August 10, 2022 Looking at all of the very large convolutional neural networks…

2 years ago

Reverse engineering the NTK: towards first-principles architecture design

Deep neural networks have enabled technological wonders ranging from voice recognition to machine transition to protein engineering, but their design…

2 years ago

Why do Policy Gradient Methods work so well in Cooperative MARL? Evidence from Policy Representation

In cooperative multi-agent reinforcement learning (MARL), due to its on-policy nature, policy gradient (PG) methods are typically believed to be…

2 years ago

FIGS: Attaining XGBoost-level performance with the interpretability and speed of CART

FIGS (Fast Interpretable Greedy-tree Sums): A method for building interpretable models by simultaneously growing an ensemble of decision trees in…

2 years ago

The Berkeley Crossword Solver

We recently published the Berkeley Crossword Solver (BCS), the current state of the art for solving American-style crossword puzzles. The…

2 years ago

Rethinking Human-in-the-Loop for Artificial Augmented Intelligence

Figure 1: In real-world applications, we think there exist a human-machine loop where humans and machines are mutually augmenting each…

2 years ago

Best Practices for Building the AI Development Platform in Government

By John P. Desmond, AI Trends Editor  The AI stack defined by Carnegie Mellon University is fundamental to the approach…

2 years ago