arXiv subjects

Archit Bansal

Publications and source records attributed to Archit Bansal.

4 recordsLinked to original sources

arXiv2025

Continual Learning for Singing Voice Separation with Human in the Loop Adaptation

Deep learning-based works for singing voice separation have performed exceptionally well in the recent past. However, most of these works do not focus on allowing users to interact with the model to improve performance. This can be crucial when deploying the model in real-world scenarios where music tracks can vary from the original training data in both genre and instruments. In this paper, we present a deep learning-based interactive continual learning framework for singing voice separation that allows users to fine-tune the vocal separation model to conform it to new target songs. We use a U-Net-based base model architecture that produces a mask for separating vocals from the spectrogram, followed by a human-in-the-loop task where the user provides feedback by marking a few false positives, i.e., regions in the extracted vocals that should have been silence. We propose two continual learning algorithms. Experiments substantiate the improvement in singing voice separation performance by the proposed algorithms over the base model in intra-dataset and inter-dataset settings.

cs.SD↗

arXiv2023

PED-ANOVA: Efficiently Quantifying Hyperparameter Importance in Arbitrary Subspaces

The recent rise in popularity of Hyperparameter Optimization (HPO) for deep learning has highlighted the role that good hyperparameter (HP) space design can play in training strong models. In turn, designing a good HP space is critically dependent on understanding the role of different HPs. This motivates research on HP Importance (HPI), e.g., with the popular method of functional ANOVA (f-ANOVA). However, the original f-ANOVA formulation is inapplicable to the subspaces most relevant to algorithm designers, such as those defined by top performance. To overcome this issue, we derive a novel formulation of f-ANOVA for arbitrary subspaces and propose an algorithm that uses Pearson divergence (PED) to enable a closed-form calculation of HPI. We demonstrate that this new algorithm, dubbed PED-ANOVA, is able to successfully identify important HPs in different subspaces while also being extremely computationally efficient.

cs.LG↗

arXiv2022

These Deals Won't Last! Longevity, Uniformity and Bias in Product Badge Assignment in E-Commerce Platforms

Product badges are ubiquitous in e-commerce platforms, acting as effective psychological triggers to nudge customers to buy specific products, boosting revenues. However, to the best of our knowledge, there has been no attempt to systematically study these badges and their several idiosyncrasies - we intend to close this gap in our current work. Specifically, we try to answer questions such as: How long does a product retain a badge on a given platform? If a product is sold on different platforms, then does it receive similar badges? How do the products that receive badges differ from those which do not, in terms of price, customer rating, etc. We collect longitudinal data from several e-commerce platforms over 45 days, and find that although most of the badges are short-lived, there are several permanent badge assignments and that too for badges meant to denote urgency or scarcity. Furthermore, it is unclear how the badge assignments are done, and we find evidence that highly-rated products are missing out on badges compared to lower quality ones. Our work calls for greater transparency in the badge assignment process to inform customers, as well as to reduce dissatisfaction among the sellers dependent on the platforms for their revenues.

cs.SI↗

arXiv2021

IITK@Detox at SemEval-2021 Task 5: Semi-Supervised Learning and Dice Loss for Toxic Spans Detection

In this work, we present our approach and findings for SemEval-2021 Task 5 - Toxic Spans Detection. The task's main aim was to identify spans to which a given text's toxicity could be attributed. The task is challenging mainly due to two constraints: the small training dataset and imbalanced class distribution. Our paper investigates two techniques, semi-supervised learning and learning with Self-Adjusting Dice Loss, for tackling these challenges. Our submitted system (ranked ninth on the leader board) consisted of an ensemble of various pre-trained Transformer Language Models trained using either of the above-proposed techniques.

cs.CL↗