Our research aims to explore new ideas and implement novel deep learning models that advance the field of artificial intelligence.

Advancing Anomaly and Churn Prediction Performance

July 26, 2022

Our researchers have applied state-of-the art techniques to improve the performance of machine learning defect detection, fraud and churn prediction. Our work on ODE-RNNs has yielded significant improvements in model performance when working with extremely large, sparse and irregular time-series data. This work has delivered millions of dollars in savings to clients.

Read More

Machine Learning and Combinatorial Optimization Problems

July 31, 2019

Perhaps the most ubiquitous of all mathematical topics that appear in industrial applications is the topic of combinatorial optimization. Combinatorial optimization is a class of problems that consists of finding an optimal object from a finite set of objects...

Read More

Generating Global Graph Structure Using Local 1D Scalar Embeddings

July 11, 2019

Random walks have been used to capture topological features (i.e. patterns of edges) of graphs since DeepWalk. It is done to obtain the graph embeddings from the samples taken along the random walks. The embeddings are a set of multivariate function on each of the nodes, S = {f(1),…, f(i),…,f(N)}. The typical number of dimensions of an embedding is in the range of [100,1000]...

Read More

Conditional GAN: A Case Study in Speech Enhancement Using Visual Cues

December 25, 2018

Recent research in GANs indicates that audio processing problems can benefit from their use. One specific problem we believe can benefit greatly from the use of GANs is speech enhancement, also known as ‘de-noising’). Speech enhancement is where relevant speech is separated out from irrelevant noises to achieve greater speech clarity...

Read More