Anubhav Gupta (@learn2phoenix) profile picture

CS Ph.D. student @UMD. Interested in Video Understanding.

College Park, MD
Joined March 2021
Photo by Anubhav Gupta (@learn2phoenix)Photo by Anubhav Gupta (@learn2phoenix)Photo by Anubhav Gupta (@learn2phoenix)Photo by Anubhav Gupta (@learn2phoenix)
We introduce LEIA, a novel approach for representing 3D articulated objects by learning view-invariant latent embeddings for different articulation states. Visit our poster #299 at @eccvconf #ECCV2024 in Milan on Oct 3rd (Thursday), 10:30am. Project Page: archana1998.github.io/leia/
4
5
7
1,888
Anubhav Gupta retweeted
🖼️How to measure whether a style exists in a generated image? If you want to know more about it, drop by our poster tomorrow at #ECCV2024 ! We also released model checkpoints and the training data! ✨ ⏰: Oct 1 (Tuesday) 10.30 session 📍- Poster # 111 💻 - somepago.github.io/csd.html #diffusion #GenAI
✨ Can we detect style in the generated images? Our recent work takes a step towards understanding this question. We train a style-focused vision feature extractor built on top of CLIP, which we call a Contrastive Style Descriptior (CSD). paper: arxiv.org/abs/2404.01292 Style similarity on CSD vectors > 0.8 more or less indicates the existence of style. And similarity less than < 0.5 indicates no style. In this figure, we can see, some artists' styles (Greg Rutkwoski, Amano) are clearly removed in the training of SD 2.1, while they are still present in SD 1.4 and Midjourney. On the other hand, we can see a strong presence of style of artists such as Afremov, Seurat, Mucha in all 3 models. #diffusion #MachineLearning 1/8 🧶
1
12
2
50
14,599
Anubhav Gupta retweeted
Measuring Style Similarity in Diffusion Models Generative models are now widely used by graphic designers and artists. Prior works have shown that these models remember and often replicate content from their training data during generation. Hence as their proliferation
4
41
11
201
96,958
📢Introducing ClimateLearn, a new PyTorch library for accessing climate datasets, state-of-the-art ML models, and high quality training and visualization pipelines. Blog: aditya-grover-group.github.i… Docs: climatelearn.readthedocs.io/ Quickstart Colab: colab.research.google.com/dr… 🧵 (1/n)
17
254
21
1,207
177,701
Anubhav Gupta retweeted
Grad students working on their papers while the world is on fire
6
29
12
476
Anubhav Gupta retweeted
Survey on Large Scale Neural Network Training A nice summary of strategies and insights on large scale neural network training. Great read for ML engineers interested in training efficient DNNs. arxiv.org/abs/2202.10435
2
91
3
387
Anubhav Gupta retweeted
Annotated (with notes) implementation of Rotary Positional Embeddings (RoPE) 📝 Code/notes nn.labml.ai/transformers/rop… It introduces a new form of positional encodings for transformers that encodes position in a relative form. Models seem to converge faster with RoPE. (1/4) 🧵👇
1
16
96
Transformer Memory as a Differentiable Search Index Shows that that information retrieval can be done with a single Transformer, in which all information about the corpus is encoded in the parameters of the model. arxiv.org/abs/2202.06991
3
53
5
341
We are releasing MOTSynth, the largest dataset for Pedestrian Detection and Tracking. Expect to obtain better real-world performances by training your model on this data only. It's synthetic, but it's for real! lnkd.in/d-NvqiX4
5
80
10
451
Anubhav Gupta retweeted
🔥The WILDS 2.0 benchmark is an oral at ICLR'22! Existing distribution shift benchmarks do not reflect the breadth of scenarios that arise in real-world applications. Wilds 2.0 extends the Wilds benchmark to include curated unlabeled data for DA research. arxiv.org/abs/2112.05090
2
29
2
155
Anubhav Gupta retweeted
We computer vision researchers rarely look at the individual data points inside our datasets. Mainly because we are too lazy and/or do not have the right tools. This needs to change. And now we have a great tool from @TensorFlow datasets team: Know Your Data. A thread.🧵(1/5)
6
126
2
687
Anubhav Gupta retweeted
A new study from @UMD_AGNR has found that some sectors, like agriculture, have barely scratched the surface to identify and address potential ethical, legal, social, and economic implications of artificial intelligence. Read the full story: go.umd.edu/qMo
2
5
I am so proud of @Waabi_ai team and everything we have achieved only 6 months on from our launch. Waabi World is the next generation of simulation technology and will be the key to unlocking self-driving at scale. We’re so excited about what’s to come! waabi.ai/waabi-world/
Today we are unveiling our revolutionary simulator—Waabi World. #WaabiWorld is the most scalable, highest fidelity closed-loop simulator ever and the key to rapidly and safely commercializing self-driving technology. waabi.ai/waabi-world/
5
25
3
162
You should always work on improving your model evaluation skills. Model evaluation is the worst taught skill in machine learning, and I believe the best way of improving is through practice. But, this paper from @rasbt is by far my favorite single written resource:
12
437
17
2,495
Replying to @Michael_J_Black
I'd also add that if you are a reviewer, and you see a co-reviewer behave in the way @Michael_J_Black describes... it's your **moral duty** to speak up and argue against him/her. That's likely the most important part of the post-rebuttal discussion phase.
2
1
33
Anubhav Gupta retweeted
Meeting senior PhD students in grad student visit days
36
8
425
Anomaly detection methods compute an anomaly score A(x), and in research, we measure their effectiveness using AUC for the binary decision "anomaly" vs. "not anomaly". But in applications, we need to choose a threshold tau. How can we set tau without having labeled anomalies? 1/
4
44
3
300
This is a big deal… awesome work from the team!!
Introducing #AlphaCode: a system that can compete at average human level in competitive coding competitions like @codeforces. An exciting leap in AI problem-solving capabilities, combining many advances in machine learning! Read more: dpmd.ai/Alpha-Code 1/
32
237
22
1,937
Anubhav Gupta retweeted
Six #UMD alums have been named in @Forbes 2022 '30 Under 30' List. today.umd.edu/briefs/6-terps…
2022 is a critical year for the self-driving industry—I’m excited to share a few thoughts on what the industry needs to do in order to make the dream of commercial self-driving cars a reality linkedin.com/pulse/how-we-ge… #autonomousdriving #SelfDrivingCars
3
18
1
82