Relational Contextual Bandits in real world user interactions Contextual bandits algorithms have become essential in real-world user interaction problems in recent years. However,these algorithms represent context as attribute value representation, which makes them infeasible for real-world domains like social networks which are inherently relational.We proposeRelational Boosted Bandits(RB2), a contextualbandits algorithm for relational domains based on relational boosted trees. RB2enables us to learn interpretable and explainable models due to the more descriptive nature of therelational representation. We empirically demonstrate the effectiveness and interpretability of RB2 on tasks such as linkprediction, relational classification, and recommendation.
Correlated discrete data generation using adversarial training Generative Adversarial Networks (GAN) have shown great promise in tasks like synthetic image generation, image inpainting, style transfer, and anomaly detection. However, generating discrete data is a challenge. This project presents an adversarial training based correlated discrete data (CDD) generation model, with a detailed approach for conditional CDD generation. Results are presented over two datasets; job-seeking candidates skill set (private dataset) and MNIST (public dataset). This project proves that the new model performs better than the existing model, as former leverages the inherent correlation in the data, while the latter overlooks correlation. Project webpage: https://sites.google.com/view/seasau
Human capital management system Finding meaningful and fulfilling work and finding right talent for a given job is a challenging Human Capital Management (HCM) problem. Seeking meaningful and fulfilling work and finding the right talent for a given job is a classical Human Capital Management (HCM) problem. This project creates a stateless scalable architecture for an automated HCM system, where algorithms that use Machine Learning techniques aid in clustering and categorizing job postings and candidate profiles. While Natural Language Processing based algorithm is used for feature extraction, a ranking algorithm that uses semantic web technologies provides accurate recommendations.