Akhila Yerukola

🌟 Thrilled to share an exciting milestone! 🌟 I’m delighted to announce that my research paper, a part of my MTech thesis at IIT Hyderabad, has been published in one of the reputed journals! 🎉 The study explores whether gamma-ray bursts (GRBs), some of the brightest explosions in the universe, can serve as reliable tools to understand our cosmos. Using innovative techniques like MCMC parameter estimation and Gaussian process regression, and leveraging data from quasars as distance markers, we examined the reliability of GRBs as cosmological probes. The results? While GRBs are fascinating phenomena, the high variability in their observable relationships suggests they might not be precise tools for unraveling cosmological mysteries—at least not independently. This journey has been both challenging and rewarding, blending cutting-edge methods with a deep curiosity about the universe. I’m grateful to my mentors, collaborators, and everyone who supported me along the way. 🔗 Paper link - https://lnkd.in/dzN9cnt8 🔗 arxiv link - https://lnkd.in/dWMKpbXK 🔗 GitHub repo link - https://lnkd.in/d3DY4hg6 #publications #AI #DataScience #Machinelearning #Paper #astrophysics

74

10 Comments

Sundar Pichai's AI Vision: Insights from CMU's Presidential Lecture Series On Wednesday 09/18, I attended a lecture by Sundar Pichai, CEO of Google and Alphabet Inc., at Carnegie Mellon University's Presidential Lecture Series hosted by Farnam Jahanian on "The AI Platform Shift and the Opportunity Ahead." Fun fact: Sundar Pichai boots up every morning with Indian Masala Chai. No wonder Google's server farms are tea-ming with activity! Looks like the secret to tech success is just a matcha luck and a whole latte innovation. ☕🖥️ Key Takeaways: AI as a Platform Shift- Sundar emphasized AI as the most profound platform shift of our time, promising an explosion of knowledge and productivity. 📈🧠 Exciting Projects- 1. Gemini: A multi-modal AI system 2. Project Astra: AI that analyzes images and answers questions about them 3. Google's innovations in deep learning architectures like transformers are pushing the boundaries of what's possible. AI and Employment- Sundar believes AI will create more jobs after a period of short-term adjustment, allowing humans to focus on creative work. 💼🎨 Responsible AI Deployment- Sundar stressed ethical AI development: - Watermarking deepfakes to prevent misuse - Prioritizing sustainability Energy and AI Synergy- AI could potentially lead to breakthroughs in sustainable energy production and efficiency. ⚡🔋 Democratizing AI- Reduced AI model costs are making AI more accessible, aligning with Google's mission. Advice to Students- Sundar emphasized finding one's passion: "Take time to figure out what your heart would enjoy doing. If you do what you enjoy, you're far more likely to do well." 💡💬💭 #AI #CMU #Innovation #ResponsibleAI #FutureOfTech #DeepLearning #QuantumComputing #Google #TechForGood #FutureOfWork #Sustainability #PlatformShift

13

Thrilled to announce that I presented my cutting-edge research on Brain-Computer Interface for Emotion Classification at the Graduate Research Expo at Northeastern University! 🧠💻🔬 As an MS Data Analytics Scholar and Google Certified Python Expert, I had the privilege of showcasing my work alongside PhD candidates, engaging with brilliant minds in the field of neurotechnology and machine learning. My research introduces a novel Hybrid Ensemble Stacking Model (HESM) for EEG Classification, leveraging network analysis and advanced hyperparameter tuning techniques. Key innovations of our model: • Developed a computationally efficient HESM that outperforms traditional deep learning methods in EEG-based emotion classification • Implemented an adaptive sliding window approach to capture temporal dynamics in EEG signals • Utilized grid search and cross-validation for robust hyperparameter optimization • Integrated diverse base classifiers (SVM, Linear SVC, MLP, Logistic Regression) to enhance generalization and model interpretability Our HESM addresses critical challenges in BCI technology: - Individual variability in emotional EEG signatures - Complex, high-dimensional nature of EEG data - Real-time processing requirements for practical applications This research has significant implications for adaptive e-learning systems, potentially revolutionizing personalized education through emotion-aware content delivery. I'm grateful for the opportunity to contribute to the advancement of BCI technology and its real-world applications. Special thanks to my advisor, Prof. Sivarit Sultornsanee for their invaluable guidance. Excited to continue pushing the boundaries of EEG signal processing and emotion recognition! Let's connect if you're interested in neurotechnology, machine learning, or adaptive learning systems. Link to video presentation: https://lnkd.in/ehqKgKZs #GraduateResearch #BrainComputerInterface #MachineLearning #EEG #EmotionRecognition #HyperparameterTuning #EnsembleLearning #AdaptiveLearning #NeuroTech #DataScience #AI #NortheasternUniversity

33

2 Comments

📢 Thrilled to share that my research on using Artificial Intelligence and Large Language Models (LLMs) for detecting COVID-19 symptoms from social media posts has been published in IEEE BIBM'23! In collaboration with Keyuan Jiang, Gordon R. Bernard and Valli Devendra, we investigated the power of pretrained LLMs like GPT-3.5 and GPT-4 to identify day-based COVID-19 symptom mentions from Twitter. This research is crucial as it helps to track health progression and symptom patterns based on self-reported data, providing insights into the spread and impact of the pandemic over time. Our findings demonstrated that these models, without any fine-tuning or supervision, achieved impressive performance in recognizing symptoms and matching them to specific days, significantly outperforming other tools like Google Bard. The results have far-reaching implications for health monitoring and public health strategies, especially in tracking disease progression through informal reports on platforms like Twitter. I'm excited to continue building on this work and explore further applications of AI in healthcare. You can check out the paper here: [https://lnkd.in/dmkvT8aV]. Looking forward to future collaborations and advancements in this field! #AI #BigData #COVID19 #Research #Pandemics #HealthcareInnovation

51

6 Comments

I'm excited to announce the successful completion of one of Northeastern's premier courses, 'High Parallel Machine Learning and Artificial Intelligence'. Professor Handan Liu curated this course with a focus on unraveling the intricacies of parallel processing, particularly in executing complex models on Discovery utilizing 28 CPU cores and 4 GPUs. The course journey was nothing short of exhilarating, where learning seamlessly intertwined with enjoyment. The excitement peaked during our project phase, centered around 'Lung and Colon Cancer Detection' employing the ResNet 50 model. Despite its pre-trained nature, mastering the utilization of code for parallel execution and attaining peak performance proved to be an enlightening endeavor. And, as per the result part we could train our model in 30 seconds using the parallel processing techniques and still maintain the accuracy to 93% overall. My project partner, Devki Patel, was also an integral part of this exhilarating journey, and together, we made significant strides in both our collaborative work and learning endeavors. Our github repository is shared in comments. Our happiness doubled when we received commendable feedback from Professor. #datascience #dataloading #modelparallel #performance #learning

61

4 Comments

I wonder why high-impact research papers on AI are not published by Indian institutes but mainly by the US or China. If you're an alumnus of an Indian tech university or a researcher (either in India or abroad), I'd love to know about your perspective on this matter. #research #ai #artificialintelligence #machinelearning #datascience #ml #softwareengineering #iit #iiit #tech #technology

2

Over last few months, I am increasingly leveraging LLMs to write code. From 10% to nearly 78% currently (based on lines of code) of my code is from LLMs. I have my own pipeline to fine tuned LLMs to augment my needs. Observations - 1. I am producing more code --> MORE IMPACT (approx 2.3X) 2. I am reading more code --> MORE MAINTAINABLE CODE 3. I am learning more as now I can start from concept and expand out ---> MORE LEARNING 4. I am incorporating best practices --> BETTER CODE 5. Lesser typing --> LOWER REPETITIVE STRAIN INJURY #AIBUDDY

76

8 Comments

🌟 Understanding the H-1B Cap-Exempt Visa: Key Considerations 🌟 When it comes to the H-1B cap-exempt visa, skipping the H-1B lottery may seem like a good alternative. However, here are a few important factors to consider before applying to qualifying educational institutions, non-profit entities, or research organizations: Advantages: 1. With a cap-exempt visa, you can avoid the lottery and significantly improve your chances of obtaining work authorization. 2. H-1B cap-exempt visa applications are accepted at any time of year. 3. Cap-exempt petitions are generally thought to have a higher chance of being approved (as long as all qualifying requirements are met). Disadvantages: 1. To transition from a cap-exempt job to a cap-subject job, you will likely need to transition your visa. 2. Cap-exempt jobs are not as widely available as cap-subject jobs. 3. Salaries might be lower compared to cap-subject roles. Sources: https://lnkd.in/eWuxCd5p https://lnkd.in/e3jmPpJF

20

1 Comment

We all thought the ultimate goal of JEE was to secure a seat in IIT—but we couldn’t have been more wrong. Almost all of us spent our last 2 years of school life preparing for JEE, alongside some folks like me who took 3 years to clear JEE, but most of us failed to achieve the success associated with it, I'm not talking about a seat in a top tier IIT, but The mindset of hard work and pushing the limits! 3.5 years post JEE, I have now realized what that phase was all about and why people who excelled in that phase tend to better in life. Moving beyond numbers and equations, it tried to teach us 👉 Committing to an Objective 👉 Working and Fighting with a Problem Statement 👉 Managing and Juggling Priorities I failed to understand and implicate them in my life, even after almost completing college, but I did retain a small fraction of it, Managing Expectations Post my JEE failure in 2020, I have started to set my expectations as per my effort rather than the facilities or past records, and yes I was in Delulu that I am good in Physics, and only that will help me achieve a decent enough rank. Well, those days are gone. Now, I look forward to what I have got in my luck, Carrer Wise. What did you retain from your JEE phase? #JEE #IIT #JEEphase #Mindset

58

3 Comments

Excited to share that our paper "ZeroHAR: Sensor Context Augments Zero-Shot Wearable Action Recognition" got accepted at AAAI 2025 (Main Technical Track)! Traditional wearable Human Activity Recognition (wHAR) systems fail to generalize to diverse human motions. Existing Zero Shot Learning methods for wHAR focus solely on augmenting labels, such as representing them as attribute matrices, images, videos, or text. We propose our ZeroHAR model architecture, that enhances time-series zero-shot learning for Human Activity Recognition by augmenting motion data with spatial and bio-mechanical sensor context features. We propose a two-stage training framework for ZSL on Wearable HAR: integrating human activity time-series data with sensor context, followed by activity recognition. Our model is generalizable across diverse human motions, with 262% improvement in zero-shot accuracy across 18 benchmark HAR datasets. Huge thanks to my brilliant co-authors Ranak Roy Chowdhury, Ameya Panse, Xiyuan Zhang, Diyan Teng, Rashmi Kulkarni, Dezhi Hong, Professor Rajesh Gupta and Prof. Jingbo Shang for their invaluable contribution and support. #AAAI2025 #MachineLearning #WearableTechnology #HumanActionRecognition #HAR

154

15 Comments

After a long break we are back with another blog and this time we are working on reversing a prefix of a word. Check out my post at https://lnkd.in/gctyzhkX

6

Hello everyone! As I am wrapping up my final year of PhD in computer science at UCSD, I would greatly appreciate any leads for industry research and academic roles. To anyone reading this-- please consider sharing my post to help me reach more people. I am passionate about healthcare and accessibility but open to exploring other domains. Please let me know if you need any references or examples of my prior work! #humancenteredai #hcai #responsibleai #ethicalai #aisafety #aialignment

77

🎉 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 Comments

🏆 Thrilled to share that my team secured 1st place at MIT Open, TechTatva'24! 🏆 MIT Open, the flagship coding contest of TechTatva, pushed us through three intense rounds of competitive programming in the ACM-ICPC style. Despite having only two team members—myself and Pranav Mathur—we gave it our all: Round 1: An online HackerRank round where 40 teams qualified out of 200. Round 2 (2 Hours) : An offline challenge requiring teamwork like never before! One member on the keyboard, another at the screen, while the third solved a bonus question—except we had no third member! Yet, we solved 4/4 questions blind and ranked 2nd despite missing the bonus. Final Round (1 hour) : A thrilling head-to-head battle where we solved 3 out of 9 questions in a lockout-type round to claim victory. I’m incredibly proud of our teamwork and determination, and this win is a testament to our problem-solving abilities under pressure. A huge thanks to ACM Manipal for hosting this incredible event! #TechTatva24 #MITOpen #CompetitiveProgramming #CodingChallenge #ProblemSolving #Teamwork #ACM_ICPC

40

4 Comments

I'm excited to share a project I've been working on with my team member Md Imtiaz Hasan Bhuiyan and Mikail Karim : an Equation Solver implemented in C++! which is our academic Project Assignment but we tried to implement by C++ Preprocessor. This project is designed to solve both linear and non-linear equations using various numerical methods. It's an excellent resource for students and professionals in scientific computing, machine learning, and related fields. And it is under improvement. Key features: - Solves linear equations using Gaussian Elimination, LU Decomposition, Jacobi Method, and Gauss-Seidel Method - Solves non-linear equations using Bisection Method, Newton-Raphson Method, Secant Method, and Fixed Point Iteration - Modular design for easy understanding and expansion - Performance measurement for comparing different methods We've created comprehensive documentation to help users understand and utilize our solver. You can find it here: https://lnkd.in/gV5HP8hT This project serves as a practical application of numerical methods, crucial for anyone studying scientific computing or machine learning. It provides hands-on experience with algorithm implementation and optimization. We welcome contributions! If you're interested in expanding the project, consider implementing optimization methods or additional numerical techniques. Special thanks to Imtiaz Hasan for his valuable contributions to the documentation. Check out our GitHub repository .https://lnkd.in/g6x88SPf to explore the code and contribute to the project.

11

Hi Connections, I am actively seeking a Summer 2025 PhD Research Internship position (US based) in cross-disciplinary research, where I can apply my expertise in machine learning—particularly geometric generative modeling and graph-based neural networks—to tackle challenging problems in Physics, Material Science, Biology, and other innovative fields. My passion lies in harnessing advanced computational techniques to simulate and analyze complex systems, bridging the gap between data-driven models and scientific inquiry. I am particularly interested in opportunities that allow me to collaborate with multidisciplinary teams spanning biology, physics, and material science to address complex challenges within these domains. By leveraging graph-based learning algotithms and geometric generative modeling, I aim to enhance our ability to model dynamic interactions, predict structural outcomes, and optimize configurations. These techniques hold the promise of transforming how we understand and manipulate biological systems, simulate physical phenomena, and design innovative materials, ultimately contributing to significant scientific and technological advancements. I am eager to join a team that values innovation and collaboration, where I can contribute my skills in graph-based modeling and generative techniques to drive forward-thinking projects and achieve impactful results. If your organization is looking for a dedicated researcher passionate about applying machine learning to solve intricate problems, I would welcome the opportunity to connect and discuss how I can contribute to your initiatives. You can find more about me and my research in my website : http://ashakya.com.np #GeometricModeling #GraphLearning #MolecularDynamics #ComputationalPhysics #MaterialScience #ComputationalBiology

34

Looking forward to sharing my thoughts on this very relevant topic of Data Science in the Age of GenAI in the #DSAA panel discussion

32

2 Comments

🚀 Discover the Cutting-Edge of AI: Agentic AI! 🚀 Today we were talking about Agentic AI in 𝕏 spaces with Priyanka Kamath and 100 Girls In Gen AI | SAAS (7K+ Community), and I figured why not write a short article about it so that people can learn the basics in 5 minutes! 🔍 What is Agentic AI? Agentic AI refers to AI systems designed to act independently, gather and analyze data using tools, and make informed decisions. These intelligent agents are transforming industries, from autonomous vehicles to advanced robotics and beyond. 🔗 Read the Full Article and Stay Ahead of the Curve! https://lnkd.in/d_EmQmD9 📢 Don’t miss out on understanding the AI revolution that’s just around the corner. Your future self will thank you! #AI #AgenticAI #Innovation #FutureTech #ArtificialIntelligence #TechTrends #LinkedInArticles

14

4 Comments