Himangi Mittal

I am a second year Master of Science in Robotics (MSR) student in the Robotics Institute (RI) at Carnegie Mellon University (CMU), working with Prof. Abhinav Gupta and collaborating with Prof. Pedro Morgado at UW-Madison. Previously, I worked as a Research Assistant at CMU with Prof. David Held at the R-Pad Lab, in collaboration with Pittsburgh-based autonomous driving company, Argo AI.

During my Masters at CMU, I am working on self-supervised representation learning methods for multimodal audio-visual videos. Previously, as a RA at CMU, I worked on self-supervised algorithms for 3D LiDAR point clouds. I did my bachelors from Jaypee Institute of Information Technology, Noida, India where I worked with Dr. Anuja Arora.

Email  /  CV  /  Google Scholar  /  Twitter  /  Github  /  Linkedin

• Jan 2023 : Teaching Assistant for 16-825: Learning for 3D Vision (being taught by Prof. Shubham Tulsiani).
• Sep 2022 : Paper accepted at NeurIPS 2022!!
• Oct 2021 : Paper accepted at BMVC 2021 (Oral).
• Apr 2021 - Dec 2021: I will be serving as a reviewer for ICCV 2021, AAAI 2022, WACV 2022, and CVPR 2022.
• Aug 2021: Journal paper accepted in PAA (in collaboration with Robert Bosch, India).
• Feb 2021: Accepted as a MSR student at CMU for Fall 2021.
• July 2020: Presented a short paper at RSS Workshop on Self-Supervised Robot Learning 2020.
• Feb 2020: Paper accepted at CVPR 2020 (Oral).

I am interested in devising self-supervised algorithms, multi-modal machine learning, video representation learning, generative models, point clouds and autonomous driving.

We propose a self-supervised algorithm to learn representations from untrimmed, egocentric videos containing audible interactions. Our method uses the audio signals in two unique ways: (1) to identify moments in time that are conducive to better self-supervised learning and (2) to learn representations that focus on the visual state changes caused by audible interactions.

A self-supervised method to complete the incomplete, partial point clouds for real-world settings like LiDAR where ground truth complete point cloud annotations are unavailable. We achieve this via inpainting where a region of the point cloud is removed and the network is trained to complete this removed region.

A method for multi-modal depression detection using audio, video, and textual modalities using LSTMs. This work leverages emotions to detect an early indication of depression.

A method of training scene flow that uses two self-supervised losses, based on nearest neighbors and cycle consistency. These self-supervised losses allow us to train our method on large unlabeled autonomous driving datasets.

Predicted action and spatial relationships in images between objects detected by YOLO, then combining VGG-Net based visual features and Word2Vec based semantic features.

A method to capture the anomalous behavior in a social network based on degree, betweenness, and closeness of graph nodes using Graph Neural Networks (GNN) in Keras.

Presents trajectory analysis of spatio-temporal graph nodes using DeepWalk algorithm in NetworkX (Python) for classification and detecting changing points of interest using SVMs.

• Teaching Assistant for 16-825: Learning for 3D Vision (Spring 2023).
• Reviewer Service: ICCV 2021, AAAI 2022, WACV 2022, CVPR 2022, CVPR 2023 (Emergency reviewer also).
• Volunteer at NeurIPS 2022 High School Outreach Program.
• Mentor at CMU AI Undergraduate Mentoring Program (Fall 2022, Spring 2023).
• Mentor at Spring 2023 CMU Research Mixer for undergraduate students organized by DPAC Undergraduate Research Working Group.

Source Code