Shaona Ghosh

Despite the ubiquity of mobile and wearable text messaging applications, the problem of keyboard text decoding is not tackled sufficiently in the light of the enormous success of the deep learning Recurrent Neural Network (RNN) and Convolutional Neural Networks (CNN) for natural language understanding. In particular, considering that the keyboard decoders should operate on devices with memory and processor resource constraints, makes it challenging to deploy industrial scale deep neural network (DNN) models. In this talk, we will cover a sequence-to-sequence neural attention network system for automatic text correction and completion. Given an erroneous sequence, our model encodes character level hidden representations and then decodes the revised sequence thus enabling auto-correction and completion. Unlike traditional language models that learn from billions of words, our corpus size is only 12 million words; an order of magnitude smaller. The memory footprint of our learnt model for inference and prediction is also an order of magnitude smaller than the conventional language model based text decoders. We report baseline performance for neural keyboard decoders in such resource constrained domain.

Shaona is a researcher in Machine Learning and NLP at Apple Inc. Previously, she was a postdoc at the Department of Engineering, University of Cambridge where she worked on developing deep learning sequence-to-sequence algorithms for prediction and auto-correction on keyboard decoders. Before that she was a postdoc in Machine Learning at the NVIDIA GPU Center of Excellence OeRC, University of Oxford. She has a PhD in Machine Learning from University of Southampton, UK. She was the Area Chair of Women in Machine Learning Workshop, 2017 and has been a reviewer at NIPS, Machine Learning Journal among others.

She has worked with and contributed to United Nations policy making for sustainable data development approaches using machine learning and for bridging the digital gender divide in AI. As a finalist of IET Young Woman Engineer in UK, 2017, she has been featured in Cosmopolitan UK. She has been selected as a ambassador for the Year of Engineering, with UK Government, Department of Transport and the IET. Her work in Cambridge was shortlisted for WISE Tech Innovation Award, 2017. Her work at HP Labs led to the merger of different business units within HP and led to a multi-national multi-year commercialization project. She has been nominated for the Telegraph top 50 Women in Engineering. She was twice awarded by Samsung Electronics for her work on innovative healthcare by using mobile phones as health sensors and predicting abnormalities using machine learning. She will be awarded the Hind Rattan (Jewel of India) award in January, 2018.

Buttontwitter Buttonlinkedin
This website uses cookies to ensure you get the best experience. Learn more