GNOME: Generating Negotiations through Open-Domain Mapping of Exchanges
Deshpande, D., Sinha, S., Kumar, A., Pal, D. & May, J. (2024). GNOME: Generating Negotiations through Open-Domain Mapping of Exchanges. ArXiv, abs/2406.10764.
CV
Education
- M.S. in Computer Science, University of Southern California, 2024
- B.S. in Computer Engineering, Mumbai University, 2022
Work experience
- August 2024 - Present: Research Engineer
- Patronus AI
- Duties include: Research on robust and automated LLM evaluations, and creation of agentic reasoning benchmarks
- August 2022 - May 2024: Research Assistant
- Information Sciences Institute
- Duties include: Conducting novel research on argumentation, and logical and commonsense reasoning in dialogue systems and LLMs.
- Advisors: Filip Ilievski, Fred Morstatter, Jonathan May
Skills
- Languages: Advanced Python, CUDA, C, Java, SQL, JavaScript, HTML/CSS
- Machine Learning: Pytorch, Tensorflow, Keras, JAX, Hugging Face, Sklearn, XgBoost, Pandas, NumPy, Matplotlib
- Deployment: Docker, Flask, Django, RestAPI, FastAPI
- Developer Tools: Git, Docker, Google Cloud Platform, Amazon AWS, Microsoft Azure
- Database Management: MongoDB, MySQL, PostgreSQL, SQLite
Publications
Contextualizing Argument Quality Assessment with Relevant Knowledge
Darshan Deshpande, Zhivar Sourati, Filip Ilievski, and Fred Morstatter. 2024. Contextualizing Argument Quality Assessment with Relevant Knowledge. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers), pages 316–326, Mexico City, Mexico. Association for Computational Linguistics.
Robust Text Classification: Analyzing Prototype-Based Networks
Sourati, Z., Deshpande, D., Ilievski, F., Gashteovski, K., & Saralajew, S. (2023). Robust Text Classification: Analyzing Prototype-Based Networks. ArXiv, abs/2311.06647.
Robust and explainable identification of logical fallacies in natural language arguments
Sourati, Z., Venkatesh, V. P. P., Deshpande, D., Rawlani, H., Ilievski, F., Sandlin, H., & Mermoud, A. (2023). Robust and explainable identification of logical fallacies in natural language arguments. Knowledge-Based Systems, 266, 110418. https://doi.org/10.1016/j.knosys.2023.110418
Talks
Diffusing Away from GANs and Transformers
Talk at TFUG Meetup, 2022, Mumbai, India
Projects and Reports
- Exploring Diffusion Models with JAX - Python, JAX, TPU (June 2022)
- Explores the mathematics and code of discrete diffusion models, their effectiveness, applicability, and limitations.
- Review Classification using Active Learning - Tensorflow, Keras, GPU (Oct. 2021)
- Demonstrates the effectiveness of active learning in text classification tasks by implementing a ratio-based sampling approach, improving performance and suitability for industrial applications.
- The Reality Behind the Optimization of Imaginary Variables - I and II - Research (Jun 2021 – Oct 2021)
- Leverages complex-valued neural networks to enhance representations by experimenting with traditional mappings and more efficient complex vector techniques.
- Explores circularity and holomorphicity constraints in complex-valued neural networks, showcasing the effectiveness of linear and widely linear networks for image-denoising.
- MelGAN based Spectrogram Inversion using Feature Matching - Reproducibility (Sep. 2021)
- Showcases a non-autoregressive model for spectral inversion utilizing a feature-matching discriminator, highlighting fast inference on dynamic inputs.
Awards and Grants
- Google OSS Expert Award (2022): For releasing open-source articles with code for exploring diffusion models.
- Papers With Code Top Contributor Award (2022): For open-source and reproducibility contributions with JAX.
- Google Research Grant (2021): A $5,000 grant to investigate the robustness of ML-based steganography methods.