Petar Veličković

X-CNN: Cross-modal convolutional neural networks for sparse datasets

In this paper we propose cross-modal convolutional neural networks (X-CNNs), a novel biologically inspired type of CNN architectures, treating gradient descent-specialised CNNs as individual units of processing in a larger-scale network topology, while allowing for unconstrained information flow and/or weight sharing between analogous hidden layers of the network—thus generalising the already well-established concept of neural network ensembles (where information typically may flow only between the output layers of the individual networks). The constituent networks are individually designed to learn the output function on their own subset of the input data, after which cross-connections between them are introduced after each pooling operation to periodically allow for information exchange between them. This injection of knowledge into a model (by prior partition of the input data through domain knowledge or unsupervised methods) is expected to yield greatest returns in sparse data environments, which are typically less suitable for training CNNs. For evaluation purposes, we have compared a standard four-layer CNN as well as a sophisticated FitNet4 architecture against their cross-modal variants on the CIFAR-10 and CIFAR-100 datasets with differing percentages of the training data being removed, and find that at lower levels of data availability, the X-CNNs significantly outperform their baselines (typically providing a 2–6% benefit, depending on the dataset size and whether data augmentation is used), while still maintaining an edge on all of the full dataset tests.

Cross-modal Recurrent Models for Weight Objective Prediction from Multimodal Time-series Data

We analyse multimodal time-series data corresponding to weight, sleep and steps measurements. We focus on predicting whether a user will successfully achieve his/her weight objective. For this, we design several deep long short-term memory (LSTM) architectures, including a novel cross-modal LSTM (X-LSTM), and demonstrate their superiority over baseline approaches. The X-LSTM improves parameter efficiency by processing each modality separately and allowing for information flow between them by way of recurrent cross-connections. We present a general hyperparameter optimisation technique for X-LSTMs, which allows us to significantly improve on the LSTM and a prior state-of-the-art cross-modal approach, using a comparable number of parameters. Finally, we visualise the models predictions, revealing implications about latent variables in this task.

Using deep data augmentation training to address software and hardware heterogeneities in wearable and smartphone sensing devices

A small variation in mobile hardware and software can potentially cause a significant heterogeneity or variation in the sensor data each device collects. For example, the microphone and accelerometer sensors on different devices can respond very differently to the same audio or motion phenomena. Other factors, like the instantaneous computational load on a smartphone, can cause key behavior like sensor sampling rates to fluctuate, further polluting the data. When sensing devices are deployed in unconstrained and real-world conditions, examples of sharply lower classification accuracy are observed due to what is collectively known as the sensing system heterogeneity. In this work, we take an unconventional approach and argue against solving individual forms of heterogeneity, e.g., improving OS behavior, or the quality/uniformity of components. Instead, we propose and build classifiers that themselves are more tolerant of these variations by leveraging deep learning and a data-augmented training process. Neither augmentation nor deep learning has previously been attempted to cope with sensor heterogeneity. We systematically investigate how these two machine learning methodologies can be adapted to solve such problems, and identify when and where they are able to be successful. We find that our proposed approach is able to reduce classifier errors on an average by 9% and 17% for a range of inertial-and audio-based mobile classification tasks.

Parapred: antibody paratope prediction using convolutional and recurrent neural networks

Motivation Antibodies play essential roles in the immune system of vertebrates and are powerful tools in research and diagnostics. While hypervariable regions of antibodies, which are responsible for binding, can be readily identified from their amino acid sequence, it remains challenging to accurately pinpoint which amino acids will be in contact with the antigen (the paratope). Results In this work, we present a sequence-based probabilistic machine learning algorithm for paratope prediction, named Parapred. Parapred uses a deep-learning architecture to leverage features from both local residue neighbourhoods and across the entire sequence. The method significantly improves on the current state-of-the-art methodology, and only requires a stretch of amino acid sequence corresponding to a hypervariable region as an input, without any information about the antigen. We further show that our predictions can be used to improve both speed and accuracy of a rigid docking algorithm. Availability and implementation The Parapred method is freely available as a webserver at http://www-mvsoftware.ch.cam.ac.uk/and for download at https://github.com/eliberis/parapred. Supplementary information Supplementary information is available at Bioinformatics online.

Automatic Inference of Cross-Modal Connection Topologies for X-CNNs

This paper introduces a way to learn cross-modal convolutional neural network (X-CNN) architectures from a base convolutional network (CNN) and the training data to reduce the design cost and enable applying cross-modal networks in sparse data environments. Two approaches for building X-CNNs are presented. The base approach learns the topology in a data-driven manner, by using measurements performed on the base CNN and supplied data. The iterative approach performs further optimisation of the topology through a combined learning procedure, simultaneously learning the topology and training the network. The approaches were evaluated agains examples of hand-designed X-CNNs and their base variants, showing superior performance and, in some cases, gaining an additional 9% of accuracy. From further considerations, we conclude that the presented methodology takes less time than any manual approach would, whilst also significantly reducing the design complexity. The application of the methods is fully automated and implemented in Xsertion library (Code is publicly available at https://github.com/karazijal/xsertion).

Scaling health analytics to millions without compromising privacy using deep distributed behavior models

People are naturally sensitive to the sharing of their health data collected by various connected consumer devices (e.g., smart scales, sleep trackers) with third parties. However, sharing this data to compute aggregate statistics and comparisons is a basic building block for a range of medical studies based on large-scale consumer devices; such studies have the potential to transform how we study disease and behavior. Furthermore, informing users as to how their health measurements and activities compare with friends, demographic peers and globally has been shown to be a powerful tool for behavior change and management in individuals. While experienced organizations can safely perform aggregate user health analysis, there is a significant need for new privacy-preserving mechanisms that enable people to engage in the same way even with untrusted third parties (e.g., small/recently established organizations). In this work, we propose a new approach to this problem grounded in the use of deep distributed behavior models. These are discriminative deep learning models that can approximate the calculation of various aggregate functions. Models are bootstrapped with training data from a modestly sized cohort and then distributed directly to personal devices to estimate, for example, how the user (perhaps in terms of daily step counts) ranks/compares to various demographics ranges (like age and sex). Critically, the users own data now never has to leave the device. We validate this method using a 1.2M-user 22-month dataset that spans body-weight, sleep hours and step counts collected by devices from Nokia Digital Health - Withings. Experiments show our framework remains accurate for a range of commonly used statistical aggregate functions. This result opens a powerful new paradigm for privacy-preserving analytics under which user data largely remains on personal devices, overcoming a variety of potential privacy risks.

Paratope Prediction using Convolutional and Recurrent Neural Networks

Antibodies play an essential role in the immune system of vertebrates and are vital tools in research and diagnostics. While hypervariable regions of antibodies, which are responsible for binding, can be readily identified from their amino acid sequence, it remains challenging to accurately pinpoint which amino acids will be in contact with the antigen (the paratope). In this work, we present a sequence-based probabilistic machine learning algorithm for paratope prediction, named Parapred. Parapred uses a deep-learning architecture to leverage features from both local residue neighbourhoods and across the entire sequence. The method outperforms the current state-of-the-art methodology, and only requires a stretch of amino acid sequence corresponding to a hypervariable region as an input, without any information about the antigen. We further show that our predictions can be used to improve both speed and accuracy of a rigid docking algorithm. The Parapred method is freely available at https://github.com/eliberis/parapred for download.

Muxstep: an open-source C ++ multiplex HMM library for making inferences on multiple data types.

MOTIVATION With the development of experimental methods and technology, we are able to reliably gain access to data in larger quantities, dimensions and types. This has great potential for the improvement of machine learning (as the learning algorithms have access to a larger space of information). However, conventional machine learning approaches used thus far on single-dimensional data inputs are unlikely to be expressive enough to accurately model the problem in higher dimensions; in fact, it should generally be most suitable to represent our underlying models as some form of complex networksng;nsio with nontrivial topological features. As the first step in establishing such a trend, we present MUXSTEP: , an open-source library utilising multiplex networks for the purposes of binary classification on multiple data types. The library is designed to be used out-of-the-box for developing models based on the multiplex network framework, as well as easily modifiable to suit problem modelling needs that may differ significantly from the default approach described. AVAILABILITY AND IMPLEMENTATION The full source code is available on GitHub: https://github.com/PetarV-/muxstep CONTACT petar.velickovic@cl.cam.ac.uk SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.