# Devi Parikh

Devi Parikh is an American computer scientist specializing in computer vision and natural language processing. She is a research director at Meta and has contributed to visual question answering and text-to-video AI models.

Devi Parikh is an American computer scientist known for her research in computer vision and natural language processing. She has held academic positions at Virginia Tech and Georgia Tech, and as of 2022 she serves as a research director at Meta, where she leads efforts in multimodal AI systems.

Parikh earned her PhD in Electrical and Computer Engineering at [Carnegie Mellon University](https://www.wikiprompt.org/wiki/carnegie-mellon-university). Her early work focused on visual recognition and scene understanding, but she later expanded into combining visual and textual information, a field that bridges [computer vision](https://www.wikiprompt.org/wiki/computer-vision) and [natural language processing](https://www.wikiprompt.org/wiki/natural-language-processing). Her contributions have been recognized with several prestigious awards, including the 2017 IJCAI Computers and Thought Award and the 2011 ICCV Best-Paper Award (Marr Prize).

## Visual Question Answering

In 2015, while at Virginia Tech, Parikh and her students developed a dataset and framework for Visual Question Answering (VQA), a task where an AI system answers questions about images, such as "Is this a vegetarian pizza?" The VQA dataset, which pairs images with human-generated questions and answers, became a benchmark for evaluating multimodal AI models. As of 2022, it has been used to evaluate over 30 AI models, influencing subsequent research in [deep learning](https://www.wikiprompt.org/wiki/deep-learning) and [multimodal learning](https://www.wikiprompt.org/wiki/multimodal-learning). This work highlighted the need for models to reason about both visual content and linguistic context, a challenge that remains central to [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) research.

## Conversational AI and ParlAI

In 2017, Parikh published ParlAI, an open-source platform for training and evaluating conversational AI agents. ParlAI integrates tasks from dialogue, question answering, and vision, allowing researchers to test models across multiple domains. The platform supports a range of [machine learning](https://www.wikiprompt.org/wiki/machine-learning) frameworks and has been used to develop agents that can hold coherent conversations while referencing visual inputs. This contribution was part of a broader effort to advance [large language models](https://www.wikiprompt.org/wiki/large-language-model) and their application to interactive systems.

## Generative AI and Dance Generation

In 2020, Parikh developed an AI system that generates dance moves in sync with music. This project, which combined [sequence-to-sequence](https://www.wikiprompt.org/wiki/sequence-to-sequence) models with audio and motion data, demonstrated the potential of generative AI in creative domains. The system learned to map musical features to choreography, producing movements that matched the rhythm and style of the input song. This work contributed to the growing field of [generative AI](https://www.wikiprompt.org/wiki/generative-ai), which also includes text-to-image and text-to-video models.

## Make-a-Video and Text-to-Video Models

In 2022, Parikh and a team at Meta developed Make-a-Video, a text-to-video AI model based on the diffusion algorithm. The model generates short video clips from textual descriptions, leveraging advances in [diffusion models](https://www.wikiprompt.org/wiki/diffusion-models) and [transformers](https://www.wikiprompt.org/wiki/transformer). Make-a-Video builds on prior work in text-to-image generation, extending it to the temporal dimension by learning to predict sequences of frames. This innovation has implications for content creation, virtual environments, and human-computer interaction, and it represents a significant step in [deep learning](https://www.wikiprompt.org/wiki/deep-learning) research.

## Awards and Recognition

Parikh's research has earned her multiple honors. In 2011, she received the Marr Prize at the International Conference on Computer Vision (ICCV), one of the top awards in computer vision, for her paper on visual attributes. In 2017, she was awarded the IJCAI Computers and Thought Award, which recognizes young researchers in artificial intelligence. These awards reflect her impact on both theoretical and applied aspects of AI, from foundational vision algorithms to practical systems that integrate language and vision.

## External links

- [Wikipedia: Devi Parikh](https://en.wikipedia.org/wiki/Devi_Parikh)

---
Source: https://www.wikiprompt.org/wiki/devi-parikh
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-10-10T05:55:36.655286+00:00