1 | Ayush Maheshwari

TL;DR: AMAL learns instance-specific weights using meta-learning to optimally combine auxiliary losses in supervised learning. Provides significant gains in knowledge distillation and rule denoising. Published at AAAI 2023.

Durga S, Ayush Maheshwari, Pradeep Shenoy, Prathosh AP, Ganesh Ramakrishnan

TL;DR: Multi-domain benchmark with 218K sentences for post-OCR correction in Sanskrit. Best model achieves 23% improvement over OCR output. Dataset spans 30 books across astronomy, medicine, and mathematics. Published at EMNLP 2022 Findings.

Ayush Maheshwari, Nikhil Singh, Amrith Krishna, Ganesh Ramakrishnan

TL;DR: Proposes robust aggregation of noisy labeling functions for semi-supervised data programming. Improves weak supervision quality by learning to weight unreliable labels. Published at ACL 2022 Findings.

Ayush Maheshwari, Krishnateja Killamsetty, Ganesh Ramakrishnan, Rishabh Iyer, Marina Danilevsky, Lucian Popa

TL;DR: Augments unsupervised constituency parsing with linguistic rules to improve performance. Combines data-driven and rule-based approaches for better syntactic parsing. Published at ACL 2021 Findings.

Atul Sahay, Anshul Nasery, Ayush Maheshwari, Ganesh Ramakrishnan, Rishabh Iyer

TL;DR: SPEAR combines semi-supervised learning with data programming to improve noisy labeling functions. Significantly outperforms state-of-the-art on seven datasets by jointly learning from rules and labeled data. Published at ACL 2021 Findings.

Ayush Maheshwari, Oishik Chatterjee, Krishnateja Killamsetty, Ganesh Ramakrishnan, Rishabh Iyer

TL;DR: Jointly learns classifier parameters and hyperbolic label embeddings for hierarchical multi-label classification without assuming known label hierarchy. Achieves state-of-the-art results by capturing hierarchical structure in hyperbolic space. Published at EACL 2021.

Soumya Chatterjee, Ayush Maheshwari, Ganesh Ramakrishnan, Saketha Nath Jagaralpudi

TL;DR: Unsupervised learning of explainable parse trees to improve model generalization in NLP tasks. Published at IJCNN 2021.

Atul Sahay, Ayush Maheshwari, Ritesh Kumar, Ganesh Ramakrishnan, Manjesh Kumar Hanawal, Kavi Arya

TL;DR: Rule-augmented sequence labeling approach for event extraction that handles tail/rare event types. Presented at StarAI Workshop (AAAI 2020).

Ayush Maheshwari, Hrishikesh Patel, Nandan Rathod, Ritesh Kumar, Ganesh Ramakrishnan, Pushpak Bhattacharyya

TL;DR: Robust end-to-end OCR systems for reading license plates and street signs in challenging real-world conditions. Practical application of computer vision for automated text recognition. Published at ICDAR 2019.

Rohit Saluja, Ayush Maheshwari, Ganesh Ramakrishnan, Parag Chaudhuri, Mark Carman

TL;DR: Uses GANs for dynamic network embedding to capture temporal evolution of networks. Presented at NeurIPS Graph Representation Learning Workshop 2019.

Ayush Maheshwari, Ayush Goyal, Manjesh Kumar Hanawal, Ganesh Ramakrishnan