PeopleofSuperintelligence

Researcher

Pranav Shyam

Pranav Shyam co-authored machine-learning papers from 2017 to 2024, and one of them, Language Models are Few-Shot Learners, was named among the NeurIPS 2020 best paper winners. He co-wrote Attentive Recurrent Comparators in 2017 and Model-Based Active Exploration in 2018, two papers on one-shot classification and reinforcement learning. In 2020 he was listed with OpenAI as an author of the GPT-3 paper, which showed that scaling up language models greatly improves few-shot performance. He co-wrote GLIDE, a text-guided diffusion model, in 2021 and Text and Code Embeddings in 2022. His most recent dated credit is the 2024 Gemini 1.5 report from Google.

Timeline

  1. March 2017

    Attentive recurrent comparators

    He co-writes Attentive Recurrent Comparators with Shubham Gupta and Ambedkar Dukkipati, a model for one-shot classification that uses only pixel information. The paper is submitted to arXiv in March 2017.

  2. October 2018

    Model-based active exploration

    He co-writes Model-Based Active Exploration with Wojciech Jaśkowski and Faustino Gomez. The reinforcement learning paper, submitted to arXiv in October 2018, introduces an algorithm that plans to observe novel events.

  3. December 2019

    Upside-down reinforcement learning

    He co-writes Training Agents using Upside-Down Reinforcement Learning with Rupesh Kumar Srivastava, Filipe Mutz, Wojciech Jaśkowski, and Jürgen Schmidhuber. Submitted to arXiv in December 2019, the paper describes a method for learning to act using only supervised learning.

  4. 2020

    Co-authoring GPT-3

    He is listed with OpenAI as an author of Language Models are Few-Shot Learners, which trains GPT-3, an autoregressive language model with 175 billion parameters. The paper, submitted to arXiv in May 2020, shows that scaling up language models greatly improves few-shot performance.

  5. 2020

    A NeurIPS 2020 best paper award

    NeurIPS includes Language Models are Few-Shot Learners among its 2020 best paper award winners.

  6. December 2021

    The GLIDE diffusion model

    He is one of the authors of GLIDE, a text-guided diffusion model for photorealistic image generation and editing. The paper is submitted to arXiv in December 2021.

  7. January 2022

    Text and code embeddings

    He is one of the authors of Text and Code Embeddings by Contrastive Pre-Training, submitted to arXiv in January 2022. The paper shows that contrastive pre-training at scale yields high quality vector representations of text and code.

  8. March 2024

    The Gemini 1.5 report

    He is among the authors of Gemini 1.5, a report from the Gemini Team at Google on multimodal models that recall information from millions of tokens of context. The report is submitted to arXiv in March 2024.

Sources and credits

Photo: We have no published photo of Pranav Shyam yet. Send us a free-licensed photo.

  1. [1703.00767] Attentive Recurrent Comparators arxiv.org. Checked 15 October 2026.
  2. Attentive Recurrent Comparators proceedings.mlr.press. Checked 15 October 2026.
  3. [1810.12162] Model-Based Active Exploration arxiv.org. Checked 15 October 2026.
  4. [1912.02877] Training Agents using Upside-Down Reinforcement Learning arxiv.org. Checked 15 October 2026.
  5. Announcing the NeurIPS 2020 award recipients – NeurIPS Blog blog.neurips.cc. Checked 15 October 2026.
  6. [2005.14165] Language Models are Few-Shot Learners arxiv.org. Checked 15 October 2026.
  7. [2112.10741] GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models arxiv.org. Checked 15 October 2026.
  8. [2201.10005] Text and Code Embeddings by Contrastive Pre-Training arxiv.org. Checked 15 October 2026.
  9. [2403.05530] Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context arxiv.org. Checked 15 October 2026.