I will be attending #NeurIPS2022 in person this week. Feel free to DM me if you are around and want to talk about self-supervised learning, multi-modal understanding, or anything over coffee. I’d love to chat 🤖😁
Follow
Click to Follow HimangiMittal
Himangi Mittal
@HimangiMittal
MS in Robotics (MSR) student at Carnegie Mellon University working in Vision, Robotics, and Self-Supervised Learning.
Himangi Mittal’s Tweets
Can we learn meaningful representations from interaction-rich, untrimmed, and multi-modal streams of audio-visual, egocentric data?
👉 Check out our #NeurIPS2022 work in Poster Session 2, 4PM - 6PM CST, Hall J # 215.
Quote Tweet
Excited to introduce our #NeurIPS2022 paper w/ Pedro Morgado, @unnatjain2010, Abhinav Gupta. We propose a method to learn meaningful representations from interaction-rich, untrimmed, multi-modal, egocentric data.
Paper: arxiv.org/abs/2209.13583
Code: github.com/HimangiM/RepLAI
Show this thread
GIF
2
4
Excited to introduce our #NeurIPS2022 paper w/ Pedro Morgado, , Abhinav Gupta. We propose a method to learn meaningful representations from interaction-rich, untrimmed, multi-modal, egocentric data.
Paper: arxiv.org/abs/2209.13583
Code: github.com/HimangiM/RepLAI
GIF
3
8
40
Show this thread
Thank you so much for tweeting about our work!
Our open-source code is also available here: github.com/HimangiM/RepLAI
#NeurIPS2022
3
📢📢🔥✨HIRING! My team at #Microsoft is hiring for full-time data & applied scientist and researcher positions for Multi-modal Content Understanding. Interested people in Vision/NLP/Speech/ML can DM me and/or apply here. 😀
careers.microsoft.com/us/en/job/1461
3
20
93
Really excited to share that my first submission at NeurIPS got accepted 🎉🎉🎉🎉 Details coming soon… #NeurIPS2022
2
29
I will present our CVPR work on "3D Reconstruction of Generic Objects Held in Hand" at Booth 129b this afternoon. Looking forward to chatting with you.
Quote Tweet
Check out @yufei_ye’s upcoming CVPR work, which shows how to reconstruct generic hand-held objects from a single RGB image. See the project page (judyye.github.io/ihoi/) for the paper, code, Colab demo to try your own images, and a video explaining the approach.
GIF
1
5
63
Show this thread
Excited to share our #CVPR2022 paper, a collaboration of & , that achieves SOTA on Online Action Detection on all benchmarks!🌟🔥
Check out our poster on Jun 24, 1430-1700, Session 4.2, ID 119. Will be in-person w/ . Drop by & say hi!👋😁
6
24
Join us for a fantastic program of thought leaders & rising stars 🌟 at #CVPR2022! 🤩
Sunday June 19 2022, 9:00-18:00 (Central Time)
#CVPR Virtual & Room 227, #NewOrleans Convention Center
Program: cvmi-workshop.github.io/program.html
#computervision #microscopy #drugdiscovery #AIforHeath
1
15
28
Please visit our #ICRA2022 Oral session (10:05AM, Room 119) and poster (Room 109-126) on May 25 to discuss Self-supervised Transparent Liquid Segmentation for Robotic Pouring.
For visualizations & videos, please see: sites.google.com/view/transpare
Quote Tweet
How can we learn to segment transparent liquids without manual labels?
Excited to share our paper which uses image translation to segment transparent liquids such as water for robotic pouring.
To appear at #ICRA2022 @CMU_Robotics
sites.google.com/view/transpare
(1/6)
Show this thread
GIF
3
6
How can we learn to segment transparent liquids without manual labels?
Excited to share our paper which uses image translation to segment transparent liquids such as water for robotic pouring.
To appear at #ICRA2022
sites.google.com/view/transpare
(1/6)
GIF
10
35
155
Show this thread
During "Spring" Break :P
Quote Tweet
Yesterday | Today
: @LwalkerCMU
: @LwalkerCMU
read image description
ALT
read image description
ALT
2
Interested in when neural nets can outperform classical methods for solving PDEs? Join us at our #NeurIPS2021 poster today at 11.30am - 1.00 pm ET if you wanna know more!
Poster: neurips.cc/virtual/2021/p
Paper: arxiv.org/abs/2103.02138
Quote Tweet
Want to learn about the neural net complexity of solutions to elliptic PDEs whose coefficients are small neural nets? Come talk to @__tm__157 , me and @zacharylipton at our poster tomorrow (Fri), 11:30a-1p EST:
neurips.cc/virtual/2021/p
Thread summarizing the paper below.
1
4
21
Starting at 14:30 (GMT/UTC) the second oral session with a general theme on 3D. See them in the live stream:
bmvc2021-virtualconference.com/conference/liv
Free to watch all, but registrations join RocketChat to ask questions.
2
8
Link to our work can be found here: bmvc2021-virtualconference.com/assets/papers/
1
Show this thread
Can we complete a partial point cloud w/o any ground truth information in real-world LiDAR data?
Check out our work on Self-Supervised Point Cloud Completion via Inpainting at #BMVC2021 in Oral Session 2 & Poster Session 1, November 22, 9:30-10:30 AM EST. ,
2
2
14
Show this thread
Don’t underestimate the power of a small step.
8
203
1,521
Quote Tweet
Excited to share our work from @MSFTResearch and @OregonState, to be presented at @ICCV_2021 as Oral. @mitts1910 #yeyu @fuxinli2 #meichen twitter.com/_akhaliq/statu…
Show this thread
2
2
Unsupervised Few-Shot Action Recognition via Action-Appearance Aligned Meta-Adaptation
pdf: arxiv.org/pdf/2109.15317
abs: arxiv.org/abs/2109.15317
project page: metauvfs.github.io
GIF
1
13
57
In about 3 weeks universities will be in session again. Many universities (like my own) want to pretend that things will be back to normal. The buildings and classrooms and quads will all be there and look the same. The routines of commuting to classes will be the same… 1/7
10
341
1,074
Show this thread
Join us for a fabulous & super-relevant program
Workshop on Computer Vision for Microscopy Image Analysis
Friday June 25 2021 12:30-20:30 (Eastern)
Zoom: lnkd.in/g_WHK9k
Program: lnkd.in/gaZ7B7H
#computervision #cvpr2021 #microscopy #drugdiscovery #pathology
12
6
I am thrilled to share that I have received the CAREER Award for "Self-supervised Representation Learning for Deformable Object Manipulation"! This research will build on insights from our work: SoftGym sites.google.com/view/softgym/h and Cloth Grasping sites.google.com/view/cloth-seg
GIF
13
13
296
Accepted to IROS 2020! #IROS2020
Jianing Qian*, Thomas Weng*, Brian Okorn, Luxin Zhang, David Held
Cloth Region Segmentation for Robust Grasp Selection
sites.google.com/view/cloth-seg
The following media includes potentially sensitive content. Change settings
View
1
8
CMU Method Makes More Data Available for Training Self-Driving Cars
2
2
A new method developed by unlocks a mountain of data that is now available to better train self-driving cars. This will help make sense of lidar sensor data, improving safety on the road. bit.ly/2BgRZmN
GIF
7
17
Happening today! #CVPR2020
Quote Tweet
Just Go With the Flow: Self-Supervised Scene Flow Estimation
Authors: Himangi Mittal, Brian Okorn, David Held
June 18 @ 11:00 AM - 1:00 PM PDT
Keywords: Scene Flow,Self-Supervised,Point Clouds,KITTI,nuScenes,Autonomous Driving #CVPR2020
1
10
Join us at #CVPR2020 from 11:00 am - 1:00 pm (PST) on Thursday 06/18 and again from 11:00pm - 1:00am (+1) PST to learn more about our work on "Just Go with the Flow: Self-Supervised Scene Flow Estimation".
Quote Tweet
Check out our #CVPR2020 oral, "Just Go with the Flow: Self-Supervised Scene Flow Estimation" which uses two self-supervised losses to learn scene flow on unannotated datasets!
Project: just-go-with-the-flow.github.io
Paper: arxiv.org/abs/1912.00497
Show this thread
GIF
1
6
Just Go With the Flow: Self-Supervised Scene Flow Estimation
Authors: Himangi Mittal, Brian Okorn, David Held
June 18 @ 11:00 AM - 1:00 PM PDT
Keywords: Scene Flow,Self-Supervised,Point Clouds,KITTI,nuScenes,Autonomous Driving #CVPR2020
1
5
13
Overwhelmed by finding optimal hyper-parameters of a CNN for a task?
Check out our HyperSTAR CVPR paper, a new method that:
☑️ “Understands” a task
☑️ Recommends hyper-parameters for it
☑️ Accelerates a hyper-parameter optimization algorithm
👉 arxiv.org/abs/2005.10524 #CVPR2020
22
78
Special thanks to all the wonderful presenters and attendees for making the 1st Neural Architecture Search for Computer Vision in the Wild Workshop at @wacv2020 a huge success! Please visit nasfw20.github.io for slides and workshop highlights.
3
12
Accepted to #CVPR2020! Just Go with the Flow: Self-Supervised Scene Flow Estimation. Himangi Mittal, Brian Okorn, David Held. New method to learn scene flow without direct supervision!
6
11