Le projet THINK

Developers can now leverage model caching through the OpenVINO™ Execution Provider for ONNX Runtime

Neuroevolution-Enhanced Multi-Objective Optimization (NEMO) for Mixed-Precision Quantization delivers state-of-the-art compute speedups and memory improvements for artificial intelligence (AI) applications.

Intel Labs has created a novel framework for producing a class of parameter- and compute-efficient models called AttentionLite, which leverages recent advances in self-attention as a substitute for convolutions.

Our latest work, presented recently at the 2021 International Conference on Learning Representations (ICLR), forces a deep network to memorize some of the training examples by randomly changing their labels.

This blog will outline the third class in this classification and its promising role in supporting machine understanding, context-based decision making, and other aspects of higher machine intelligence.

In this post, I introduce a metric for estimating the complexity level of your dataset and task, and I describe how to utilize it to optimize distillation performance.

Model distillation is a powerful pruning technique, and in many use cases, it yields significant speedup and memory size reduction.

We describe a scalable framework that combines Deep RL with genetic algorithms to search in extremely large combinatorial spaces to solve a critical memory allocation problem in hardware.

In this blog, I intend to explore this method further and investigate other test classification datasets and sub-tasks in an effort to duplicate these results.

Machine learning requires us to have existing data — not the data our application will use when we run it, but data to learn from.