Measuring dataset similarity using optimal transport

Is FashionMNIST, a dataset of images of clothing items labeled by category, more similar to MNIST or to USPS, both of which are classification datasets of handwritten digits? This is a pretty hard question to answer, but the solution could have an impact on various aspects of machine learning. For example, it could change how practitioners augment a particular dataset to improve the transferring of models across domains or how they select a dataset to pretrain on, especially in scenarios […]

Read more

Claraprint: a chord and melody based fingerprint for western classical music cover detection

Cover song detection has been an active field in the Music Information Retrieval (MIR) community during the past decades. Most of the research community focused in solving it for a wide range of music genres with diverse characteristics… Western classical music, a genre heavily based on the recording of “cover songs”, or musical works, represents a large heritage, offering immediate application for an efficient fingerprint algorithm. We propose an engineering approach for retrieving a cover song from a reference database […]

Read more

Weakly Supervised Learning of Nuanced Frames for Analyzing Polarization in News Media

In this paper we suggest a minimally-supervised approach for identifying nuanced frames in news article coverage of politically divisive topics. We suggest to break the broad policy frames suggested by Boydstun et al., 2014 into fine-grained subframes which can capture differences in political ideology in a better way… We evaluate the suggested subframes and their embedding, learned using minimal supervision, over three topics, namely, immigration, gun-control and abortion. We demonstrate the ability of the subframes to capture ideological differences and […]

Read more

Integration of Clinical Criteria into the Training of Deep Models: Application to Glucose Prediction for Diabetic People

Standard objective functions used during the training of neural-network-based predictive models do not consider clinical criteria, leading to models that are not necessarily clinically acceptable. In this study, we look at this problem from the perspective of the forecasting of future glucose values for diabetic people… In this study, we propose the coherent mean squared glycemic error (gcMSE) loss function. It penalizes the model during its training not only of the prediction errors, but also on the predicted variation errors […]

Read more

Big Announcement: 3 Free Certificate Courses in Data Science and Machine Learning by Analytics Vidhya!

An Unmissable Opportunity to Earn your Data Science Certificate Picture this – you are given the opportunity to take a high-quality course on a data science or machine learning topic(s) free of cost. And as the icing on an already delicious offering, you will even get a certificate upon completing the course! So not only do you get to embellish your blossoming data science skillset, you get a certificate proof of your accomplishments. Sounds too good to be true? Well, […]

Read more
1 883 884 885 886 887 910