Sumedh Koppula

🎇 On National Day, I went for a leisurely drive in San Francisco and ended up "stress-testing" a Waymo self-driving car on the road. 🚗 As an autonomous driving practitioner, who wouldn't be curious about the real-time robustness of the cutting-edge Waymo One driving system? While cruising downtown, I accidentally noticed a Waymo car tailgating me. While this isn't unusual for SF citizens, a wild, "evil" idea suddenly hit me: Why not directly "adversarially attack" the world's autonomous driving status quo? I executed an unexpected maneuver—suddenly reversing—to see how it would react. 🌟 The response was stellar! At the moment I reversed, Waymo One honked instantly—quicker than any human could—activated its hazard lights, and backed away to maintain a safe distance. This reckless move on my part served as an edge case to test the algorithm's robustness under extreme conditions and, potentially, could be a challenging training sample to enhance Waymo's future autonomous systems. 💫 Thrilled to personally experience how current cutting-edge autonomous algorithms handle rare driving behaviors—and how stable and safe Level 4 autonomy is in dealing with diverse scenarios. However, it also prompted deep reflection as an AI researcher in this field: 🤔 In an industry with little room for error, how can we ultimately avoid or minimize issues that AI fails to handle? 💡 I believe two research directions are particularly promising for achieving Level 5 level autonomy in future mobility systems: - 1️⃣ Development and deployment of vehicle-to-everything (V2X) cooperative systems (including V2V, V2I, V2P, etc). Our initial studies (e.g., V2X-ViT, ECCV'2022 arxiv.org/abs/2203.10638) show that in scenarios with severe occlusions or noise, such cooperative systems can significantly enhance the robustness of perception systems, thereby eventually improving traffic safety. - 2️⃣ Adversarial scenarios generation (including Sim-to-Real, generative modeling). Research done by my colleagues at UCLA (V2XP-ASG, ICRA'2023 https://lnkd.in/gu5nKVHD) shows that adversarial learning techniques can effectively simulate adversarial scenarios, greatly improving model robustness in complex "corner case" situations. Of course, it's often infeasible to collect such collision scenarios. 👨‍🏫 As a new Assistant Professor of CS at Texas A&M University, I will lead a group focusing on these exciting research directions, which can be a proactive approach to reducing accidents and improving safety for all. 🔥 I look forward to future collaborations with governments, academics, and companies to research and develop data and algorithms that can help enhance the safety of vulnerable road users, especially seniors and children. We envision a people-centered intelligent transportation system in the future. Interested in these topics? Let's connect and discuss further! #AutonomousVehicles #AI #MachineLearning #SmartCities #Transporation #Humanity #Mobility

144

3 Comments

https://lnkd.in/gXVrPZE One of my favorite papers. It got basically no recognition, but is really useful for discrete time quantum walks on graphs.

3

👉 Learn how to simulate pre-trained models from #PyTorch and #TensorFlow directly in #Simulink with the new co-execution Simulink blocks. https://spr.ly/6044W3drs

2

Check out this blog post on how to use the newest MATLAB functions to convert deep learning models between MATLAB, PyTorch, and TensorFlow. #deeplearning #matlab #pytorch #tensorflow

12

 A pre-print of my summer intern's recent paper, entitled “Domain decomposition-based coupling of operator inference reduced order models via the Schwarz alternating method”, is available on ArXiv here: https://lnkd.in/g8qZ94CA. The paper presents and evaluates an alternating Schwarz-based approach for coupling together subdomain-local reduced order models constructed via non-intrusive Operator Inference (OpInf) with each other and with subdomain-local full order models (FOMs).

53

Since our new 🆕 LABLES 👨‍💻 code is built on PyTorch, and since PyTorch has MPS backend support, we can run a 256x256 2D LBM lid driven cavity at Re=500,000 in 7 minutes flat 🚀 on an Apple M1 Macbook Pro 💻 ...just be sure to have your charging cable 🔌 and a power outlet handy as it drains the battery 🔋 pretty fast 🏃‍♂️...#latticeboltzmannmethod #gpucomputing

211

22 Comments

A Simple Example Using PCA for Outlier Detection https://ift.tt/oipk7r6 Improve accuracy, speed, and memory usage by performing PCA transformation before outlier detection This article continues a series related to applications of PCA (principle component analysis) for outlier detection, following Using PCA for Outlier Detection. That article described PCA itself, and introduced the two main ways we can use PCA for outlier detection: evaluating the reconstruction error, and running standard outlier detectors on the PCA-transformed space. It also gave an example of the first approach, using reconstruction error, which is straightforward to do using the PCA and KPCA detectors provided by PyOD. This article covers the second approach, where we first transform the data space using PCA and then run standard outlier detection on this. As covered in the previous article, this can in some cases lower interpretability, but it does have some surprising benefits in terms of accuracy, execution time, and memory usage. This article is also part of a larger series on outlier detection, so far covering FPOF, Counts Outlier Detector, Distance Metric Learning, Shared Nearest Neighbors, and Doping. This article also includes another excerpt from my book Outlier Detection in Python. If you’re reasonably familiar with PCA itself (as it’s used for dimensionality reduction or visualization), you can probably skip the previous article if you wish, and dive straight into this one. I will, though, very quickly review the main idea. PCA is a means to transform data (viewing data records as points in high-dimensional space) from one set of coordinates to another. If we start with a dataset (as shown below in the left pane), with 100 records and two features, then we can view the data as 100 points in 2-dimensional space. With more realistic data, we would have many more records and many more dimensions, but the same idea holds. Using PCA, we move the data to a new set of coordinates, so effectively create a new set of features describing each record. As described in the previous article, this is done by identifying orthogonal lines through the data (shown in the left pane as the blue and orange lines) that fit the data well. So, if we start with a dataset, such as is shown in the left pane below, we can apply PCA transformation to transform the data into something like is shown in the right pane. In the right pane, we show the two PCA components the data was mapped to. The components are simply named 0 and 1. Left pane: 100 data points in a dataset with two features. The blue and orange lines show orthogonal lines that may be drawn through the data to capture the location of the points well. These are used to determine the PCA transformation. Right pane: the same data after PCA transformation. We have the same 100 data points, but two new coordinates, called Component 0 and Component 1. One thing to note about PCA components is that they are...

1

🚀An outstanding paper at #ICLR2024 on improving the shortcomings of #VisionTransforms - "Vision transformer s need registers". Transformers have recently emerged as a powerful tool for learning visual representations. In this paper, we identify and characterize artifacts in feature maps of both supervised and self-supervised ViT networks. The artifacts correspond to high-norm tokens appearing during inference primarily in low-informative background areas of images, that are repurposed for internal computations. We propose a simple yet effective solution based on providing additional tokens to the input sequence of the Vision Transformer to fill that role. 🔥Kudos to the team of Timothée Darcet Read more here - https://lnkd.in/d6z8nRHk #ViT #transformers #iclr #research #deeplearning #artificialintelligence

16

🚨Paper Alert 🚨 ➡️Paper Title: Local Policies Enable Zero-shot Long Horizon Manipulation 🌟Few pointers from the paper 🤖Sim2real for robotic manipulation is difficult due to the challenges of simulating complex contacts and generating realistic task distributions. 🤖To tackle the latter problem, authors of this paper introduced “ManipGen”, which leverages a new class of policies for sim2real transfer: local policies. Locality enables a variety of appealing properties including invariances to absolute robot and object pose, skill ordering, and global scene configuration. 🤖They combined these policies with foundation models for vision, language and motion planning and demonstrate SOTA zero-shot performance of our method to Robosuite benchmark tasks in simulation (97%). 🤖They transferred their local policies from simulation to reality and observed that it can solve unseen long-horizon manipulation tasks with up to 8 stages with significant pose, object and scene configuration variation. 🤖ManipGen outperforms SOTA approaches such as SayCan, OpenVLA, LLMTrajGen and VoxPoser across 50 real-world manipulation tasks by 36%, 76%, 62% and 60% respectively. 🏢Organization: Carnegie Mellon University, Apple 🧙Paper Authors: Murtaza Dalal, Min Liu, Walter Talbott, Chen Chen, Ph.D. Deepak Pathak, Jian Zhang, Russ Salakhutdinov 1️⃣Read the Full Paper here: https://lnkd.in/gVeKy3dE 2️⃣Project Page: https://lnkd.in/gBAp8aQ5 🎥 Be sure to watch the attached Demo Video - Sound on 🔊🔊 Find this Valuable 💎 ? ♻️REPOST and teach your network something new Follow me 👣, Naveen Manwani, for the latest updates on Tech and AI-related news, insightful research papers, and exciting announcements.

23

New paper: Shared Memory Parallelism in Modern C++ and HPX https://lnkd.in/ghYcgGkQ

39

A session on "Julia Compiler Secrets - Static Analysis" by Keno Fischer and Shuhei Kadowaki The Julia compiler has evolved significantly, expanding its static analysis capabilities beyond type inference to include effect analysis, escape analysis, and optimized generics. These advancements, introduced in recent releases like Julia 1.8 and 1.11, are paving the way for faster, more efficient, and safer code. The talk delved into how these new capabilities enhance code performance and correctness, offering insights into the future of the Julia compiler. Exciting times ahead for Julia developers #JuliaCon #JuliaLang #CompilerTech #StaticAnalysis #PerformanceOptimization

75

🤖 Check out the MATLAB Deep Learning Hub on GitHub! What you'll find: 🧠 Examples and tutorials for deep learning concepts 🚀 Pretrained models to kickstart your project Get Started: 🖱️ Go to the MATLAB Deep Learning Hub on GitHub. 👀 Browse the resources. ⭐ Download, star, or fork your favorites and apply them to your projects! Click the link in our bio! #DeepLearning #MachineLearning #AI #MATLAB #GitHub #LearnToCode

1

- 'Allegro Hand' by Wonik Robotics and Meta Platforms - "Three new innovations we’re sharing today: 1. Meta Sparsh: first general-purpose encoder for vision-based tactile sensing that works across many tactile sensors and many tasks. 2. Meta Digit 360: an artificial fingertip-based tactile sensor that delivers detailed touch data with human-level precision and touch-sensing. 3. Meta Digit Plexus: a standardized platform for robotic sensor connections and interactions that enables seamless data collection, control and analysis over a single cable," says AI at Meta. “Wonik Robotics and Meta FAIR aim to introduce robotic hands to global companies, research institutes, and universities so they can continue developing robotic hand technology that is safe and helpful to humankind,” says Dr. Yonmook Park, Executive Director and the Head of Future Technology Headquarters at Wonik Robotics Co., Ltd. - Source - AI at Meta's Threads: https://lnkd.in/g2jySpq2 October 31, 2024 By Meta Platforms Meta FAIR's Blog: https://lnkd.in/gkX9ptB3 Advancing embodied AI through progress in touch perception, dexterity, and human-robot interaction October 31, 2024 By Meta Platforms - End - Yonmook Park(박연묵)

15

2 Comments

This is #icra2024 Week, and I want to tell you about a very cool paper I spotted from Waymo's research team: Stateful Tracking with Transformers for Autonomous Driving! It's a 3D Multi-Object Tracker... but with Transformers rather than Kalman Filters. It's very "edgy", and I'm doing a complete breakdown in this video. Dragomir Anguelov (Drago) Qinru Li How did I do? Did I miss anything? #deeplearning #computervision #lidar #selfdrivingcars #robotics

119

6 Comments

Amazing findings by Hugh Zhang, Michele L., Summer Yue, and others at Scale AI, addressing a critical concern regarding Large Language Models (LLMs) evaluation and the challenge of data contamination! With LLMs achieving remarkable success on various benchmarks, concerns have arisen about the extent to which this performance reflects true reasoning ability versus exposure to similar benchmark questions. In response, the paper https://lnkd.in/gttAW-g2 undertook an investigation, introducing Grade School Math 1000 (GSM1k) to closely mirror established benchmarks like GSM8k. Their findings shed light on the prevalence of overfitting, particularly among models like Mistral and Phi, which exhibit accuracy drops of up to 13% on GSM1k compared to GSM8k. However, models at the forefront, such as Gemini, Claude, and GPT, demonstrate minimal signs of overfitting. This research not only underscores the importance of robust benchmarking but also prompts critical reflections on the evaluation of LLM performance. Kudos to the team for their work! #llm #research #genai #machinelearning #ai

15

5 Comments