Subhashree Radhakrishnan

Our paper, "Adapting multilingual vision language transformers for low-resource Urdu optical character recognition (OCR)" is now available for reading in the PeerJ Computer Science journal. This was an effort on researching bilingual NLP techniques (localization, img2text and multimodal NER) for a low-resource language (Urdu in our use case) coupled with English. This paper is an extension of our FYP, alongside Mohammad Daniyal Shaiq, Musa Dildar Ahmed Cheema and Dr M. Asif Naeem. Paper here: https://lnkd.in/eC-A2r4Q

115

13 Comments

this is best place to start studying for your ML journey

3

Transformers Transformers has been a major milestone in the field of NLP and heavily used in Generative AI. There are 3 variants of transformer based models and they are 1. Encoder-only 2. Decoder-only 3. Encoder-Decoder Encoder only models: These are also called as autoencoders and these are pretrained using a technique called masked language modeling. A text with a random masked token is sent to the model to predict the masked token. For example, consider the text "If you don't stop at the sign, you will get a ticket" The training input that will be passed on the encoder model is "If you don't _____ at the sign, you will get a ticket. The model is expected to predict the token (word) "stop" These models use bi-directional representations of the input to better understand the full context of a token. Examples of encoder models: BERT Family models Decoder only models: These models are called autoregressive models and are pretrained using a technique called causal language modeling. These models predict the next token using the previous tokens. For example, consider the text "If you don't stop at the sign, you will get a ticket" The training input that will be passed on the encoder model is "If you don't ______" The model will still try to predict the word "stop", but only based on previous tokens. These models are used for generative tasks, including question answering. Examples of decoder only models: GPT Family, Falcon, LLama models Encoder decoder models: These models are called sequence to sequence models and the pretraining objective varies from model to model. For example the popular FLAN - T5 uses a consecutive multitoken masking called span corruption. For example, consider the text "If you don't stop at the sign, you will get a ticket" The training input that will be passed on the encoder model is "If you don't _____ _______ the sign, you will get a ticket. The model will try to predict the tokens "stop at" These models are good at translation tasks. Examples: T5 family In all the above explanations we just took one sentence of text. The LLMs you see in the market are trained on huge volumes of text available over the internet. #LLM #encoder #decoder #transformer #genai

3

🌟 Exciting News! 🌟 Thrilled to share that our latest paper, "On Evaluating Black-Box Explainable AI Methods for Enhancing Anomaly Detection in Autonomous Driving Systems," has been published in MDPI Sensors Journal! In this work, we introduce a comprehensive framework to assess black-box XAI techniques like SHAP and LIME for anomaly detection in autonomous vehicles. Our study evaluates these methods across six key metrics, providing valuable insights into their effectiveness and limitations. A big thank you to my amazing co-authors OSVALDO ARRECHE and Mustafa Abdallah for their hard work and collaboration. Read the full paper here: https://lnkd.in/dqkUdEQ2 #AutonomousDriving #ExplainableAI #AnomalyDetection #XAI #AI #Cybersecurity #Research

36

16 Comments

 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

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

In this episode, we discuss CYCLE-CONSISTENT LEARNING FOR JOINT LAYOUT-TO-IMAGE GENERATION AND OBJECT DETECTION by The paper's authors are listed as "Anonymous authors" since it is under double-blind review.. The paper introduces a new generation-detection cycle consistent (GDCC) learning framework that simultaneously optimizes layout-to-image generation and object detection, highlighting the inherent duality of these tasks. GDCC employs cycle losses to guide both tasks, enhancing data efficiency without requiring paired datasets, and achieves computational efficiency through novel sampling strategies while keeping inference cost unchanged. Experimental results demonstrate that GDCC improves diffusion model controllability and object detector accuracy, with plans for code release.

2

🔗 Simulate PyTorch Models with Simulink! 🚀 Engineers, unlock new possibilities by simulating PyTorch and other Python-based models directly in Simulink using co-execution blocks. This integration streamlines your workflow, enabling seamless model testing and validation. 🔍 What's Inside: >> Co-execution of Python models in Simulink << >> Enhanced simulation and validation processes << 👉 Dive into the details: https://spr.ly/6041o1cOX #DeepLearning #Simulink #Python #ModelSimulation #Engineering #MATLAB #MAthWorks #PyTorch

9

We had a great session about NVIDIA NIMs with Anshul Jindal and Mark Moyou, PhD yesterday. 🚀 From Mark: We got lots of info on inferencing with NIMs and different types of NIMs like embedding models, generative models, reranking models.. 🎨 I walked through how you can use NIMs for embedding, retrieval and generation with #Haystack 🚀 And finally, Anshul did a live coding session where he showed how you can deploy and scale your own self-hosted NIMs. If you missed it, or don't even know what I'm talking about but you wanna learn, the recording is up (adding to the comments👇) All resources, including the example repos and guides are in the description. #AI #LLM #Nvidia #opensource

35

2 Comments

simple AI paper review : When using existing machine learning models without explicit causal information, there is a risk of misinterpreting correlations as causal relationships, particularly in complex scenarios with sequential effects. This paper introduces an advanced methodology for analyzing pixel circuit design, integrating Explainable AI (XAI) techniques, specifically SHAP values, to address these challenges. By constructing a Directed Acyclic Graph (DAG) in collaboration with domain experts, the complexity of interconnections was reduced, improving causal analysis accuracy. The methodology enhances understanding of causal mechanisms in pixel circuitry, demonstrating the effectiveness of combining machine learning with XAI and gaining acceptance among experts. REF : https://lnkd.in/g6QesN8r

11

Our paper on #XR and #haptic enabled #hri (human #robot interaction) framework is available for download now... https://lnkd.in/gGYrkpVZ

26

Transformer-based LLMs are extraordinary, but their self-attention mechanism struggles with long sequences due to quadratic complexity - making inference both costly and slow. This new research introduces "Star Attention" - NVIDIA's novel algorithm that speeds up LLM inference by up to 11x while preserving 95-100% accuracy! 🚀 🔑 Key Innovation: 🔹 Splits long contexts into blocks with an anchor block 🔹 Processes context in parallel across multiple hosts 🔹 Maintains global attention capabilities 🔹 Reduces memory requirements dramatically 🍀 Technical Breakthrough: 1️⃣ Blockwise-local attention processing 2️⃣ Sequence-global attention for queries 3️⃣ Seamless integration with existing transformer models 📈 Performance Impact: 🔹 Up to 11x faster inference 🔹 Works on sequences up to 1M tokens 🔹 Minimal accuracy loss 🔹 Increasingly impressive with larger models While Star Attention shines in most scenarios, refining anchor block size and improving performance on complex multi-hop reasoning tasks are still some questions that need to be answered. Nevertheless, the trend is loud and clear: faster models that require less memory. 💡 Read more 👉 https://lnkd.in/gRkeW-jA #ai #genai #llms #transformers

7

Excited to share that our extended abstract titled "Incorporating Foundation Model Priors in Modeling Novel Objects for Robot Instruction Following in Unstructured Environments" has been accepted for presentation at the Workshop on 3D Visual Representations for Robot Manipulation at ICRA 2024! Visit our website for more details: https://lnkd.in/gcmfMJgu Workshop: https://lnkd.in/gA38HrpH 📝 Authors: Moksh Malhotra, Aman Tambi, Sandeep Zachariah, P.V.M. Rao, Rohan Paul 🙏 Special thanks to our project advisors Prof. Rohan Paul and Prof. P.V.M. Rao for their invaluable guidance and support throughout this research journey! In our paper, we tackle the challenge of acquiring object models when a robotic manipulator is sent in unstructured and unknown environments. We proposes a novel approach to utilizes foundation models with RGB-Depth images to acquire high-fidelity semantic 3D model of objects, required for robotic manipulation. We demonstrate our pipeline in simulation as well as real-world scenarios. Looking forward to discussing our research and exchanging ideas with fellow researchers and professionals at the conference. Stay tuned for more updates! #ICRA2024 #Robotics #Research #AI #3DVisualization #RobotManipulation

11

2 Comments

I'm thrilled to announce that our paper, "Cumulant Learning: Highly Accurate and Computationally Efficient Load Pattern Recognition Method for Probabilistic STLF at the LV Level", has been accepted for publication in the prestigious IEEE Transactions on Smart Grid! 🎉 This paper is a generalization of my PhD thesis and the result of a collaborative effort with my professors/mentors and co-authors: Goran Švenda, Aleksandar Erdeljan, Milan Gavric, and Darko Čapko. I’m especially proud that this paper was recognized by one of the top journals in the field of Electrical and Electronics Engineering, ranking in the top 10% according to the WoS/JCR impact factor (classified as M21a in Serbia). 🏆 This work tackles one of the biggest challenges in energy forecasting—predicting the highly variable load patterns of millions of low-voltage consumers (think about all the times we turn on and off lights, air conditioners, washing machines, and other devices at home). 💡 So, rather than trying to predict the exact behavior of every individual with overly complex models (and likely failing spectacularly), we asked ourselves, “What if we could predict the collective behavior of similar consumers and use that to estimate the future behavior of individuals with some level of probability?” 🤔 This approach still required a specially designed clustering method and a fairly deep neural network, but it paid off: our model achieved the highest average accuracy in a case study on real-world smart meter data from the UK and Australia, even outperforming more sophisticated models like graph neural networks. On top of that, it proved more robust to outliers and better at predicting load surges, with shorter execution times and lower memory consumption, making it highly suitable for managing low-voltage networks. ⚡🌍 Feel free to check out the early access link below and reach out to me for a free copy! 😉 I welcome any feedback or questions you might have! #SmartGrid #Clustering #DeepLearning #PatternRecognition #Forecasting

47

4 Comments

Interesting article about Generative AI! Subscribe to Rishabh Yadav's blog for many more!

10

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