Скачать или смотреть Smart Neural Network Optimization

Running a computer vision model is an expensive operation. You have invested a lot of resources and time building your model. Now it turns out keeping the lights on might be even more costly. And then you realise this will cost a lot of energy as well.

During this talk Steven van Blijderveen explains Convolutional Neural Networks (CNN) and different types of compression methods available for CNNs, such as quantization and pruning. Quantization refers to the process of reducing the number of bits that represent a number. In the context of neural networks, this means using lower-precision formats to represent weights and activations, which can lead to significant reductions in model size. Pruning is a compression technique that involves eliminating unnecessary connections or weights in a neural network. For example, if we imagine a neural network as a vast web of interconnected neurons, pruning can be likened to trimming off the less important connections, allowing the network to focus on the more significant ones.

Besides explaining about these types of compression, Steven will also talk about knowledge distillation, which is a technique where a compact neural network, known as the student, is trained to imitate a larger, more complex network or ensemble of networks, known as the teacher. The student network learns from the output of the teacher network rather than the raw data, enabling it to achieve comparable performance with a fraction of the resources.

Matthijs Plat explains about the solution that AIminify has built for the high energy consumption and how this can be used in the world of AI.

For presentation slides and Q&A please visit the event page at https://www.itc.nl/about-itc/centres-...