Priyanka Golia
ACM India Doctoral Dissertation Award
India - 2024
citation
For her dissertation titled "Functional synthesis via formal methods and machine learning"
Priyanka Golia Chosen as Recipient of the ACM India 2024 Doctoral Dissertation Award
The ACM India 2024 Doctoral Dissertation Award goes to Priyanka Golia for her dissertation titled "Functional Synthesis via Formal Methods and Machine Learning." Priyanka’s dissertation combines ideas from many areas—constraint sampling, machine learning, and formal methods, and based on this, she built a solver, Manthan, for synthesis of Boolean functions. Manthan is available as open-source and is currently the state-of-the-art tool for Boolean functional synthesis. Priyanka’s dissertation work was done at IIT Kanpur under the supervision of Prof. Subhajit Roy, IIT Kanpur and Prof. Kuldeep Meel of University of Toronto.
At its core, Manthan uses a data-driven approach to solve the problem. It works in four stages: (1) data generation, (2) learning an ML model, (3) translating ML model to Boolean functions, and (4) formal verification and repair. While her initial work on Manthan led to strong publications, Priyanka did not stop there. She improved Manthan via many strong algorithmic and engineering innovations: a pre-pass to identify uniquely determined Skolem functions, learning a multi-class decision tree instead of a binary classifier, and using lexicographical MaxSAT. After this, Priyanka further lifted Manthan to learning Boolean functions with explicit dependencies, addressing the well-known problem of learning Henkin functions for DQBF instances. After her stint with developing tools for functional synthesis, Priyanka explored applications for such functions, and the possibility of using similar approaches for real-world problems. Priyanka therefore receives the DDA award for her contributions to functional synthesis, building an open-source tool based on her ideas, and exploring the application of these ideas to real-world applications.
The Honorable Mention for 2024 goes to Abhijeet Awasthi for his dissertation titled "Learning Language Processing Models With Limited Labeled Data." Abhijeet’s thesis makes pioneering contributions to data-efficient NLP, including methods for collecting efficient supervision, synthetic data generation, and specialized architectures for learning with limited labeled data. These techniques have applications in question-answering over databases, grammatical-error-correction, text-to-AP! calling, and speech-recognition. Abhijeet’s dissertation work was done at IIT Bombay under the supervision of Prof. Sunita Sarawagi.
Abhijeet has made extremely valuable contributions to grammar error correction and has influenced the model that was adopted in the well-known tool, Grammarly, for error correction. His framework, PIE, obtains almost SOTA accuracy in GEC for much less computational resources, thereby addressing practical issues for real-world applications. PIE provided inference-time speed-ups of up to 1,400% relative to the then state-of-the-art Grammatical Error Correction Models from Google while also offering a slightly higher accuracy. Abhijeet has released high-quality packages that have gotten hundreds of stars on GitHub, and models that have been included in third party libraries. Almost every paper is accompanied by a public release of code and datasets. For example, the work on high-level supervision has been incorporated in three standard libraries and benchmarks: Astra (Microsoft), WRENCH (University of Washington), and Spear (IIT Bombay).
The ACM India Doctoral Dissertation Award was established in 2011. This award recognizes the best doctoral dissertation in Computer Science and related disciplines from a degree-awarding institution based in India for each academic year, running from July 1 of one year to June 30 of the following year. The ACM India Doctoral Dissertation Award is accompanied by a prize of ₹2,00,000. An Honorable Mention award, given to nomination(s), if any, that missed the award by a narrow margin, is accompanied by a prize of ₹1,00,000, that is shared among the recipient(s). The winning dissertation(s) will be published in the ACM Digital Library. Tata Consultancy Services Limited (TCS) is the founding sponsor of these awards. Please see the ACM India Doctoral Dissertation Award page for additional information on current and past winners.
Please join us in congratulating Priyanka Golia and Abhijeet Awasthi for their significant achievements.