Liza Dahiya

To(ring) or not to ring. That's the question we tried to answer. Last quarter, I along with Pranava Singhal and Harshvardhan Agarwal worked on a project to answer a simple question in Graph ML: Can adding virtual nodes to represent molecular rings improve performance on graph-level tasks? 🌟 Why do rings matter? Rings in Molecules play a role in determining their properties, like hydrophobicity, which influence solubility, and bioavailability. Inspired by this, we proposed a novel RingSAGE algorithm to explicitly tackle these structures during the message-passing stage in GNNs. 💡 Why This Matters Traditional GNNs treat molecular graphs uniformly, missing the unique properties of rings. Our approach introduces virtual nodes to capture ring-specific information. However, this leads to an increase in the number of learned parameters, so for a fair comparison, we compare our method against a variant of GIN with more parameters than our model. 🔍 What We Achieved Our model outperformed standard GNN models on the ZINC dataset, a widely used molecular drug solubility prediction benchmark. We provide a notebook to replicate these experiments, a codebase to use RingSage for your own usecase and a blog elaborating our method and results.

73

6件のコメント

Launching yaR! We built an AI pendant to help visually impaired people navigate the world. We call it yaR, and we need your help to make it real. It all started back in January at NUS Hack and Roll. Sparsh ., Shrivardhan Goenka, Yajat Gulati and I were throwing around ideas, and we landed on this: could AI help visually impaired people understand their surroundings? We weren't sure, so we asked Ajay Minocha, CFA, a visually impaired CFA at Deloitte. His response? A resounding yes - this space needs innovation. So we built a prototype. Let me tell you, it wasn't pretty. Cardboard case, boom mic, Raspberry Pi 4 - it looked like something out of a sci-fi B-movie. But you know what? It worked. We could've stopped there. Posted about it and moved on. But we couldn't shake the feeling that we'd stumbled onto something bigger. So yaR was born. Since then, it's been a wild ride. We've been working closely with the Singapore Association of the Visually Handicapped, getting real feedback from the people who'd actually use this. Our professors and mentors - Prof Mohamed M. Sabry Aly, Prof Arvind Easwaran, Prof Bo An, Prof Jingxian Wang, Assoc. Prof. Suranga Nanayakkara, Nuwan Janaka and Keat from NUS-HCI Lab, and Jovan Tan from NUS Innovation and Design Programme - they've been simply incredible, guiding us through the maze of tech and business. We've hit roadblocks. So. Many. Roadblocks. Hardware delays, bugs that made us want to tear our hair out, moments where we wondered if we were in over our heads. But we kept pushing. Four iterations of the prototype later, yaR is starting to look like something real. Now, we're taking the next big step. We're launching a Kickstarter where you can pre-order yaR v1. This isn't a finished product - it's the start of something we want to grow with your help. That's why we're also open-sourcing yaR. Whether you're a coder, a designer, or just someone with ideas, you can help shape what yaR becomes. Here's how you can get involved: 1. Kickstarter: https://lnkd.in/gyEuaXUu 2. GitHub: https://lnkd.in/gfjT8ZK9 3. Spread the word - your network might be the key to reaching someone who needs a yaR We've been lucky and are grateful to have some amazing supporters along the way: - 1517 Fund Medici Project: USD 1,000 - Microsoft for Startups: USD 25,000 in cloud credits - Engineers Innovation Challenge: SGD 5,000 - Student Entrepreneurship Program @ NTU: SGD 3,000 We're not just building a device. We're trying to change how visually impaired people experience the world. It's a big goal, and we can't do it alone. So we're asking for your help. Back us if you can, contribute if you're able, share if you're willing. Let's make yaR a reality together. Want to know more? Visit www.yar.computer We'd love to hear what you think. This is just the beginning. Join us?

199

19件のコメント

Should the current placement situation in IIT's be an excuse for T2-T3 colleges to simply shake off their responsibility of placing their students ? The job market has been absolutely shit, and there's no shame in admitting it. There's no shame in reaching out to your alumni network to ensure these emerging leaders get opportunities to flourish in life. What's shameful, however, is when you don't take cognizance of the matter and do nothing, just because asking for help can label you as WEAK in front of many. A humble request to many placement cells out there, this is not the time to exhibit your high handedness over students ; this is the time to work together to ensure people get the opportunities they deserve. This is the time to COLLABORATE. P.S if anyone is looking to get into data analytics, I have a few leads. DM me and I'll share those with you. Praying for better times ahead 🙏🏻

69

15件のコメント

🎉 Great News! 🎉 I'm thrilled to share that our(Animesh Mukherjee, Paramita Das and Ritabrata Chakraborty) paper "On the effective transfer of knowledge from English to Hindi Wikipedia" has been accepted at the COLING 2025 (Industry Track) 🌟! In this work, we propose a lightweight framework that enriches Hindi Wikipedia articles by extracting and adapting updated information from English sources, ensuring alignment with Wikipedia’s style and policies, and then translating the adapted content into Hindi to significantly improve the coverage and quality of these articles. Key Contributions include: 1️⃣ We propose a multistage approach to extract knowledge from biographical writings about popular figures, transform this text to adhere to the NPOV guidelines and incorporate it into low-resource Wikipedia articles, e.g., Hindi. 2️⃣ We manually curate 103 biographical writings relevant to corresponding Wikipedia articles as external knowledge sources, which can serve as a rich source of factual information 3️⃣ Our lightweight framework efficiently updates Hindi Wikipedia articles by adding coherent and new information. 📄 The preprint is available on ArXiv: https://lnkd.in/dQ2axSx4 We are thrilled to present this work at COLING 𝟮𝟬𝟮𝟱 and look forward to connecting with the AI research community 🌐🤝.

64

4件のコメント

🔍 Day 4 of my 10-Day PyTorch Learning Streak: Exploring Binary Classification Today’s focus was on binary classification, where the goal is to predict one of two possible outcomes—such as "yes" or "no." 🩺 Example Project: I worked on a project predicting heart disease. Using health parameters as input data, I built a PyTorch model to determine whether or not an individual has heart disease. 💡 Key Highlights from Today: 1️⃣ Data Preparation: Cleaned and preprocessed the input data to ensure it was ready for training. 2️⃣ Model Architecture: Designed a neural network optimized for binary classification tasks. 3️⃣ Loss Function & Evaluation: Used Binary Cross-Entropy Loss and tracked accuracy to evaluate the model's performance. 4️⃣ Insights Gained: Explored how to manage imbalanced datasets and refine predictions for better accuracy. 📂 Python Code Included: For those interested, the Python code for the binary classification project is included in this update for your reference. Github link : https://lnkd.in/dvqupd5e 🤝 Apologies for the Delay: Sorry for not posting this update earlier—I've been busy with studies, but I promise there’s more exciting content coming soon. Stay tuned! Binary classification is such a crucial concept in machine learning, enabling applications like disease prediction, fraud detection, and beyond. Every step in this journey continues to strengthen my PyTorch skills and confidence. 📘 Excited for the next milestone in this adventure! #PyTorch #MachineLearning #BinaryClassification #DeepLearning #DataScience #HeartDiseasePrediction

5

🚀AI-Powered Search Engine Using Arxiv, Wikipedia, and Intelligent Agents by Krish Naik Sir🌐 Introducing my latest project—an AI-powered Search Engine that integrates Arxiv and Wikipedia tools alongside advanced agent-based interaction! This tool brings together cutting-edge technologies to deliver smarter, faster, and more efficient search capabilities. 💡 Features & Capabilities: - 📚 Arxiv Integration: Access concise academic summaries and research papers. - 🌍Wikipedia Tool: Fetch accurate and reliable information from the world's largest encyclopedia. - 🔍 DuckDuckGo Search: Fast and privacy-focused web search for additional insights. - 🧠 LangChain Agents: Seamlessly manage multiple tools and provide meaningful outputs with the help of Zero-Shot ReAct agents. - ⚡Streamlit UI: A sleek and user-friendly interface for easy interaction. 🛠️ Tech Stack: - LangChain for AI-driven agent orchestration. - Streamlit for interactive web-based interfaces. - Arxiv & Wikipedia API Wrappers for custom data retrieval. - DuckDuckGo for quick search capabilities. 👉GitHub Link:https://lnkd.in/eCdFd3Ey Check out the code to understand the thought process behind each line and the seamless integration of multiple tools! 💻 Whether you're exploring Machine Learning or diving into AI Search Systems, this project showcases a powerful blend of research, innovation, and technology.

36

2件のコメント

I met Abhishek Gupta in 2022 and weirdly, we tried hiring each other, for our individual startup ideas. He was looking for someone technical who could help him code a restaurant PoS system. We didn't team up because we thought "ideas > cofounder" (classic mistake) Weeks later I met Abhishek, he said he would not do any intern but teach himself full stack dev so he could build whatever he wished. "That's gutsy! let's see what he does" - is what I thought 1 year later Abhishek had learnt complete end to end development and was in fact consulting other startups in software building and deploying. Then I remember at one point we were talking about B2B sales. Another year later, Abhishek had cracked into the toughest market in insurance, purely by his strategy of building great relationships with CXOs over WhatsApp. His journey was different from most people I've met. He decides, takes the risk and delivers. There's a lot to learn from Abhishek's journey of starting from scratch and building an intuition and then expertise in anything - that goes beyond the scope of this article too! So it was natural that he, Jayanth Krishnaprakash and I teamed up building Recontact, and that journey is still in the making. If you ever want good feedback on a business idea, he's a great guy to chat, given your ask good questions :)

93

1件のコメント

𝗧𝗵𝗲 𝗗𝗮𝘆 𝗜 𝗦𝘁𝗼𝗽𝗽𝗲𝗱 𝗛𝗮𝘁𝗶𝗻𝗴 𝗣𝗵𝘆𝘀𝗶𝗰𝘀—𝗛𝗲𝗿𝗲’𝘀 𝗪𝗵𝗮𝘁 𝗖𝗵𝗮𝗻𝗴𝗲𝗱🎢: Class 11th. JEE prep was in full swing. And there was one subject I hate with all my heart: 𝗣𝗵𝘆𝘀𝗶𝗰𝘀. 😩 I would often joke that Physics was a foreign language sent to test my patience. But something changed during my 12th year. After countless failed attempts at dodging Physics, I had a realization. It wasn’t about fighting the subject but understanding 𝘄𝗵𝘆 I hated it. And the answer was simple: 𝘐 𝘥𝘪𝘥𝘯’𝘵 𝘤𝘰𝘯𝘯𝘦𝘤𝘵 𝘸𝘪𝘵𝘩 𝘪𝘵. Once I realized this, I decided to approach Physics differently—not as an obstacle but as a puzzle waiting to be solved. 🔧 𝗧𝗵𝗲 𝘁𝘂𝗿𝗻𝗶𝗻𝗴 𝗽𝗼𝗶𝗻𝘁? All thanks to Alakh Sir! Fortunately in class 12th, I discovered his YouTube channel PW (PhysicsWallah), and everything changed. His teaching made Physics feel less like a burden and more like a challenge I could actually overcome. I started breaking down problems into smaller, more manageable pieces, asking more questions, and relating concepts to everyday life. Slowly, things started making sense. I’m not saying I became a Physics genius overnight, but I stopped seeing it as my enemy. Instead, it became a challenge I was eager to overcome. 𝗧𝗵𝗲 𝗹𝗲𝘀𝘀𝗼𝗻? Sometimes, it’s not about whether we like something, it’s about finding the right way to approach it. With the right mindset and guidance, even the toughest obstacles can be overcome. 👉 Have you ever struggled with a subject or task but found a way to turn it around? Let me know in the comments—I'd love to hear your story! 👇 #JEEPrep #PhysicsStruggles #AlakhSir #GrowthMindset #OvercomingChallenges #LearningJourney #Class12 #coding #learn #tech #linkedin #challenge

38

6件のコメント

"Community building via entrepreneurship starts with understanding you want to provide for your workers. You hire good people willing to work for the leadership, and leadership cares and rewards you for your efforts. You will have created an environment where your employee’s families can thrive with psychological safety from the concept of greed of “corporate psychopathy” that is so prevalently encouraged to entrepreneurs in US investor scene - and you will have done your job in creating generational wealth for your teams lifetime." #leadership #community #entrepreneurship

2

1件のコメント

Dear Quant Club IIT Jodhpur Students, Are you ready for today's exciting offline contest on Matiks? 🚀 Contest Format: Optiver 80 in 8 You will have 80 questions to solve in just 8 minutes! Each contest window will be open for 20 minutes window, and you can start at any time within the first 12 minutes to get the full 8 minutes. If you begin after the 12th minute, you'll only have the remaining time to complete the questions, which will be less than 8 minutes. So, plan your start time wisely! ⏱️ Today's Contest Structure: 0️⃣ First Contest (Warm-Up): This is just a practice round of 30 minute to familiarize you with the contest format and will not count toward the final results. 🔢 Rounds 1, 2, and 3: You will receive results for each round individually. Additionally, we will provide an Excel sheet where you can view the combined results of all three rounds. If you haven’t explored our platform yet, we highly recommend trying out our daily challenges to get familiar with the pattern of questions and sharpen your skills 🧠 Best of luck to everyone! We can't wait to see you tackle the challenge and achieve great results. 💪 #IITJodhpur #Matiks

99

3件のコメント

📢 Exciting News: My Paper is Published! 📢 I’m Harshit, and I’m thrilled to announce that my research paper titled "Cloud Computing Digital Image Processing through Comparative Analysis of CVIP & MATLAB Tools" has been published in the UGC-approved journal Shodhasamhita! 🎉 Publication Date: Volume-X, Issue-I (VI), Jan-June 2023 In this study, we developed the CVIP MATLAB toolbox package, a comprehensive collection of computer vision and image processing functions. The toolbox now includes 201 functions covering various aspects of image analysis and computer vision, such as arithmetic and logic, band, color, conversion, edge/line detection, image geometry, histogram, mapping, morphological operations, noise handling, objective fidelity metrics, pattern classification, distance/similarity metrics, feature extraction, normalization methods, segmentation, spatial filters, and transform filters. This toolbox aims to enhance the capabilities and efficiency of digital image processing tasks in cloud computing environments. Additionally, my Main thesis, titled "Comparative Analysis of Different Cloud Computing Security Services Offered by Microsoft Azure, Amazon Web Services & Google Cloud Platform," provides an in-depth comparison of the security services offered by these major cloud platforms. I’m incredibly grateful for the support and guidance from my professors, peers, and mentors throughout this journey. A huge thank you to everyone who has been a part of this achievement. If you’re interested in reading the full paper or learning more about my thesis, feel free to reach out or access it through [provide link or further instructions if available]. Looking forward to sharing more updates and engaging with fellow researchers in the field! #Research #AcademicPublishing #UGCJournal #Shodhasamhita #MTech #ProudMoment #ResearchPaper #AcademicAchievements #CloudComputing #Security #MicrosoftAzure #AWS #GoogleCloud

14

1件のコメント

Mental health is a real issue that deserves our attention and care. Academic stress is one of the many factors that can significantly impact mental well-being, but it should never be allowed to escalate to the point of suicide. Let's prioritize creating supportive environments where individuals feel safe seeking help and where mental health resources are readily available and destigmatized. Together, we can work towards fostering a culture of understanding, empathy, and support for those struggling with mental health challenges. Your well-being matters, and there is always hope and help available. Sharing the profiles of therapists whom I trust and if you feel like you or someone you know needs help, then defiantly reach out to them- 1- Soni Jaiswal 2- Vinamra V. 3- Ananya Pant #wellness #therapy #mentalhealth

28

2件のコメント

🌟 Just Completed the Ultimate PyTorch Course!  I'm thrilled to share that I've just completed an extensive PyTorch course from Zero To Mastery through Udemy! Here’s a snapshot of what I learned: 📚 1. PyTorch Fundamentals: Understanding tensors and how they are used in machine learning. 2. PyTorch Workflow: Steps from data to tensors to trained neural network models. 3. Neural Network Classification: Building models to classify data. 4. Computer Vision with PyTorch: Creating models for image recognition. 5. Custom Datasets: Loading and processing custom data. 6. Modular Code: Writing reusable Python scripts for machine learning. 7. Transfer Learning: Leveraging pre-trained models for new tasks. 8. Experiment Tracking: Keeping track of model performance. 9. Paper Replicating: Replicating cutting-edge research papers. 10. Model Deployment: Deploying models to the web. This particular course stood out to me for many reasons:   -Hands-On Projects: Running experiments, completing exercises, and building real-world models. -Comprehensive Coverage: From basics to advanced topics, everything is covered step-by-step. -Expert Instruction: Learning from a professional machine learning engineer with real-world experience. -Community Support: Access to an exclusive online community of students, alumni, mentors, TAs, and instructors. I'm excited to apply these skills in creating innovative AI solutions, and pursuing opportunities as a Machine Learning Engineer. The journey has just begun, and I'm ready to take on new challenges in this rapidly evolving field! #PyTorch #MachineLearning #DeepLearning #AI

25

2件のコメント

If you are attending ICPR 2024 in the next 5 days (Dec 01-05), then do visit our poster titled: "waLLMartCache: A Distributed, Multi-Tenant and Enhanced Semantic Caching System for LLMs”; my co-author Binay Gupta will be presenting it. Our system waLLMartCache advances the GPTCache framework by incorporating the following features: 1. We introduce the support for a new database Redis in GPTCache – this is used as L2 cache in our designed system. 2. Currently, GPTCache is implemented to be run on a single node which we enhance to span across multiple nodes to handle industry-scale requests and consequently, we also designed a distributed eviction manager. 3. We further create partitions for individual tenants (clients) so that these can be hosted together while maintaining semantic separations. 4. We present a decision engine (i.e., an enhanced semantic similarity checker) that decides whether to cache an LLM response based on retail business use-cases. 5. We showcase that loading FAQs (which can be set to be stored persistently in the memory) while booting the LLM cache can be a simple yet effective strategy to boost cache hits significantly. Some of these contributions have already been merged with the open-source GPTCache code. A big shout out to my other co-authors: Soumik Dasgupta, Anurag Wagh, Lalitdutt Parsai, Geet Vudata, Shally Sangal, Sohom Majumdar, Hema Rajesh, Anirban Chatterjee. #GenAI #LLMs #MachineLearning #ArtificialIntelligence #Industry #Research #Conference #Paper

116

3件のコメント