Rationalizing Neural Predictions

The study aims to overcome the missing transparency and interpretability of neural network predictions. The authors propose a modular framework which learns to extract key phrases for sentiment analysis and text classification. Although restricted to NLP problems in this study, the basic approach of extracting the most relevant features for accurate prediction might be very…

Read More

Residual Networks and Recurrent Neural Networks

Shows the equivalence of very deep Residual Networks and shallow Recurrent Neural Networks with identity shortcut mappings. Shallow Recurrent Neural Networks are closer to the biological reality of the primate visual cortex than a deep Residual Network with thousands of layers. We discuss relations between Residual Networks (ResNet), Recurrent Neural Networks (RNNs) and the primate…

Read More

Progressive neural networks

First steps in reinforcement learning towards continual learning and transfering knowledge to new domains. Still restricted to compatible source and task domains. Learning to solve complex sequences of tasks—while both leveraging transfer and avoiding catastrophic forgetting—remains a key obstacle to achieving human-level intelligence. The progressive networks approach represents a step forward in this direction: they…

Read More

AlphaGo – Mastering the game of Go with deep neural networks and tree search

Complexity too high for straightforward data-driven learning approaches Combination of supervised learning to assess board positions and reinforcement learning for probabilistic policies to restrict the space of possible moves Specialized AI, learnt skills not transferable/applicable to other contexts The game of Go has long been viewed as the most challenging of classic games for artificial intelligence…

Read More