Satya Sai Srinath Namburi’s Post

𝐄𝐱𝐜𝐢𝐭𝐢𝐧𝐠 𝐍𝐞𝐰𝐬 𝐢𝐧 𝐎𝐛𝐣𝐞𝐜𝐭 𝐃𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧: 𝐘𝐎𝐋𝐎-𝐖𝐨𝐫𝐥𝐝 Thrilled to share an innovative new development in the field of object detection - YOLO-World. Building on the efficiency and practicality established by the You Only Look Once (YOLO) series of detectors, YOLO-World brings an open-vocabulary detection capability to the table. Traditional detectors are limited by predefined and trained object categories. However, YOLO-World moves beyond these confines by incorporating vision-language modeling and pre-training on large-scale datasets: this manifests as consistent, exceptional performance in detecting an expansive range of objects in a zero-shot manner while maintaining high efficiency. The technological innovation behind this approach is the newly-proposed Re-parameterizable Vision-Language Path Aggregation Network (RepVL-PAN) combined with a region-text contrastive loss. These facilitate a more profound interaction between visual and linguistic information. The results speak for themselves: on the challenging LVIS dataset, YOLO-World achieves a 35.4 average precision (AP) at an impressive 52.0 frames per second (FPS) on V100. This achievement outperforms many state-of-the-art methods in terms of both speed and accuracy. More interestingly, the fine-tuned YOLO-World shows remarkable performance on several downstream tasks. This includes object detection and open-vocabulary instance segmentation, highlighting broad applications and potential for this technology. 🔗 https://lnkd.in/gzYp_b2w YOLO-World is a serious game-changer, introducing flexibility and scalability to object detection that was previously unattainable. Stay tuned for more developments in this space! #AI #ObjectDetection #MachineLearning #yolo #computervision #datascience #artificialintelligence #innovation #technology #visionmodeling #YOLO-World

🎉 Thrilled to announce that two papers have been accepted to #ICLR2024 📄 Paper Title: "Conditional Instrumental Variable Regression with Representation Learning for Causal Inference" 🔍 This paper studies the challenging problem of estimating causal effects from observational data, in the presence of unobserved confounders. We use a conditional IV (CIV) to relax the unconfounded instrument condition of standard IV and propose a non-linear CIV regression with Confounding Balancing Representation Learning, CBRL.CIV, for jointly eliminating the confounding bias from unobserved confounders and balancing the observed confounders, without the linearity assumption. We theoretically demonstrate the soundness of CBRL.CIV. confounders. You can find the paper on OpenReview here: https://lnkd.in/gS-FWUtz 📄 Paper Title: "Causal Inference with Conditional Front-Door Adjustment and Identifiable Variational Autoencoder" 🔍 In this paper, we relax some of the restrictions by proposing the concept of conditional front-door (CFD) adjustment and develop the theorem that guarantees the causal effect identifiability of CFD adjustment. By leveraging the ability of deep generative models, we propose CFDiVAE to learn the representation of the CFD adjustment variable directly from data with the identifiable Variational AutoEncoder and formally prove the model identifiability. You can find the paper on OpenReview here: https://lnkd.in/g_TcMvSv 👀 Looking forward to sharing my research at the conference and further discussing developments in this field.

"YALTAi introduces a game-changing approach to layout analysis in OCR and similar tasks. By leveraging object detection instead of pixel classification, it significantly enhances segmentation efficiency. The incorporation of YOLOv5 into Kraken 4.1's pipeline yields remarkable performance gains, particularly on smaller datasets. This innovation marks a pivotal shift in document digitization, promising superior extraction accuracy and noise reduction. #AI #MachineLearning #ComputerVision #YALTAi #KrakenEngine"

https://lnkd.in/dKqkVPKm This study addresses the intricate cold start challenge in movie recommendation systems (RSs) by integrating Collaborative Filtering (CF) and Singular Value Decomposition (SVD), complemented by Generative Adversarial Networks (GAN). The cold start problem arises when the system is unable to provide accurate recommendations for new users or items with limited interaction history. The addition of Content-Based (CB) filtering enhances outcome depth, while thorough data exploration and meticulous feature engineering, including extracting details like release years, enrich the dataset for a nuanced understanding of user preferences. The unique strength lies in the synergistic application of CF-SVD, GAN, and CB filtering, providing a multifaceted strategy to overcome the cold start challenge. Results, guided by strategic CB filtering, offer comprehensive insights, emphasizing the pivotal role of content in crafting accurate and personalized movie recommendations. The innovative integration establishes our approach as a robust solution, proficient in providing nuanced suggestions, especially in scenarios with limited user interaction data, thereby advancing the efficacy of movie RSs. GAN, a state-of-the-art framework, contributes significantly to our integrated RSs by leveraging a generator model to create synthetic data, refining CF outcomes. This contribution is evaluated using the MovieLens dataset, encompassing ratings and movie information.

Google DeepMind Researchers Propose a Novel Divide-and-Conquer Style Monte Carlo Tree Search (MCTS) Algorithm ‘OmegaPRM’ for Efficiently Collecting High-Quality Process Supervision Data

Google DeepMind Researchers Propose a Novel Divide-and-Conquer Style Monte Carlo Tree Search (MCTS) Algorithm 'OmegaPRM' for Efficiently Collecting High-Quality Process Supervision Data https://lnkd.in/gwb5WPS6

📢 [New blog article - Understanding Variational Inference and Data Generation] 📢 https://lnkd.in/erNXeaVM After weeks of autodidactic dedicated work, I’m excited to share my latest article, which dives into the foundations of Variational Inference (VI) as a stepping stone to understanding Variational Autoencoders (VAEs). 📈🧠 The article explores how VI, when applied to deep learning, enables the creation of generative models like the VAE, with transformative potential in data modeling and generation. My goal was to distill the complex mathematics behind VI and showcase how it leads to the innovative architectures found in VAEs today. 🧩🔍 This experience proved to me that it's possible to dive into a new topic and, within weeks, publish an article that explains the core concepts clearly. In addition, I’m thrilled to announce the upcoming start of a new project with my professors Loic Le Folgoc and Elsa Angelini at Télécom Paris, focusing on applying NVAE (Normalizing Flow Variational Autoencoder) in medical imaging. 🚀💡 To support this journey, I launched my own website, where this article and others on deep learning topics are now accessible. https://lnkd.in/ehrqgmqm Feel free to explore and connect! 🔗🌐 #AI #VAE #MedicalImaging #DeepLearning #TelecomParis

“Intervention and Conditioning in Causal Bayesian Networks” is not just a paper; it’s a revolutionary map that navigates the complex maze of causality. 🤯 What’s the Buzz? Imagine being able to predict the ripple effects of one action across a network of variables. That’s exactly what this paper achieves with its cutting-edge approach to Causal Bayesian Networks (CBNs). It’s like having a crystal ball for data scientists! 👨🔬 The Genius Behind the Scenes Kudos to the brainiacs, Sainyam Galhotra and Joseph Y. Halpern, for cracking the code on how to calculate conditional probabilities with interventions. They’ve turned what was once a head-scratcher into a no-brainer! 🚀 Why It Matters? This isn’t just academic jargon; it’s a game-changer for anyone in AI. From healthcare to finance, the implications are as vast as they are thrilling. We’re talking about more accurate predictions, smarter decisions, and AI that truly understands cause and effect. Paper: https://lnkd.in/gfypFjMG #DataScience #MachineLearning #BayesianNetworks #ai

RAG Optimization - Episode #1: Current challenges in RAG optimization largely revolve around balancing computational efficiency with the quality of generated content, while also managing the ever-growing size of knowledge bases. The primary need for optimization stems from the necessity to deliver high-quality, contextually relevant responses in real time, without incurring prohibitive computational costs or delays. At Ascendion, we focus on enhancing the speed and accuracy of information retrieval from large datasets, improving the model's ability to integrate this information seamlessly into the generative process, and ensuring the model remains up-to-date with the latest information. Let’s talk about Retrieval Efficiency through Vector Quantization. For retrieval efficiency, a pivotal component is the use of vector quantization techniques within the embedding space to speed up document retrieval. Vector quantization involves mapping high-dimensional vectors (document or query embeddings) into a finite set of vectors in a lower-dimensional space. The objective function for training a vector quantizer can be defined as minimizing the average distortion. Efficient quantization allows for rapid nearest neighbor searches in the quantized space, significantly reducing the computational cost of retrieval from large knowledge bases. Sounds interesting? Reach out to Muthu Chandra to get more insights into RAG optimization. Arun Varadarajan Prakash Balasubramanian Santhosh Mukundan Ramakrishnan JN David Larson Karthik Krishnamurthy Paul Roehrig, PhD Reshma Rahi Pandian Muneeswara C Viral Tripathi Ian Lee (he/him) Joshua Lee #RAG #Optimization #GenerativeAI #cost #quality #schedule