John Hughes·Sep 2, 2020How to Boost Emotion Recognition Performance in Speech Using Contrastive Predictive CodingIn this article I will take you through how I developed an emotion recognition system using speech as an input and then boosted performance using self-supervised representations trained with Contrastive Predictive Coding (CPC). Results have improved from a baseline of 71% to 80% accuracy when using CPC. …Machine Learning14 min read
Michaeltansinispeechmatics·Aug 12, 2020How to optimise automatic speech recognition performance using MKLNote: This blog entry contains a number of acronyms. If you start to feel lost, there is a glossary at the bottom of the page. At Speechmatics, we’re constantly looking for ways to improve the accuracy and efficiency of our language products. We do this in multiple ways: Gathering more useful data. Language packs are trained on data from a variety of sources. It’s important that this data reflects the use cases of our customers…Intel8 min read
David MacLeod·Oct 31, 2019How to Build a Streaming DataLoader with PyTorchThe release of PyTorch 1.2 brought with it a new dataset class: torch.utils.data.IterableDataset. This article provides examples of how it can be used to implement a parallel streaming DataLoader in PyTorch, as well as highlighting potential pitfalls to be aware of when using IterableDatasets for sequential data. If you want your own copy of this article, head over to Speechmatics’ website to download my eBook.Pytorch12 min read
Will Williams·Aug 15, 2019How to Deploy Machine Learning in ProductionHints and tips on how you can help your business deploy machine learning into production so that you and your customers can consume it. This article is based on my own experiences of using machine learning and applying it to speech recognition at Speechmatics. For your very own copy of…Docker18 min read
Eric Henderson·Jul 17, 2019How to Evaluate Speech Technology Providers — 4 Key ConsiderationsEvaluating speech technology providers can be difficult and frustrating. What do you measure? How do you measure it? Often the unit of measurement when evaluating speech technology providers is the word error rate (WER), but what about language coverage, speed of process, deployment options? …Machine Learning6 min read
Tom Ash·May 30, 2019How to make the most of machine learningAn introduction to machine learning — what is it and how can we use it? This is an entry-level introduction to machine learning based on a course I conducted within my organisation, Speechmatics. This blog addresses what machine learning is, why we use it, what it can do, and how we can train it. To receive your own copy of this blog, download my eBook…Machine Learning17 min read
James Dobson·May 16, 2019What do DevOps and Machine Learning Have in Common?DevOps and Machine Learning What do DevOps and machine learning have in common? Pipelines Data storage/Manipulation Orchestration Scaling Packaging All these things are in traditional microservices and what you experience for building/training a machine learning component. …Docker4 min read
Anartz Nuin·Mar 28, 2019How to write Kubernetes custom controllers in GoA Kubernetes controller is an active reconciliation process. It watches the shared state of the cluster through the API server and makes changes attempting to move the current state towards the desired state. The intention of this article is to provide an introduction on how to write Kubernetes custom controllers…Kubernetes3 min read
Tom Ash·Mar 21, 2019How to make the most of data surplusA three-part blog series addressing the reality of having a surplus of data in the increasingly data-driven world. In this series, I’ll offer possible solutions to our excess of data, from tips on how to train on a surplus of data and advice on how to reduce your data down…Machine Learning5 min read
Tom Ash·Mar 14, 2019How to make the most of data surplusA three-part blog series addressing the reality of having a surplus of data in the increasingly data-driven world. In this series, I’ll offer possible solutions to our excess of data, from tips on how to train on a surplus of data and advice on how to reduce your data down…Machine Learning5 min read
