Ankur Gaur

🤔 Interesting perspective from Infosys Chairman Nandan Nilekani: "Let Silicon Valley build LLMs; India should lead in AI use cases." At Meta's Build with AI summit, Nilekani suggested that India should focus on practical AI applications rather than competing in the LLM space. But is this the right approach? Key points to consider: • Silicon Valley giants are heavily invested in LLM development • India could focus on: - Synthetic data generation - Small language models - Real-world applications - Building robust data infrastructure The vision is for India to become the global "use case capital" for AI, emphasizing frugal innovation and practical solutions. What do you think? 💭 Should India stick to developing practical AI applications? 💭 Or should we also invest in fundamental AI research and LLM development? 💭 Could focusing solely on use cases limit our AI capabilities in the long run? Let's discuss! Share your thoughts in the comments below. 👇 #AI #DigitalIndia #LLM

30

Text chunks in retrieval may risk losing context. The below article introduces contextualizing each chunk using an LLM. The added cost and latency required for contextualizing is partially addressed through a lightweight LLM (haiku) and prompt caching. Along with other known techniques (BM25, reranking, etc.), a significant reduction in retrieval failure is achieved. Overall, a good article to read. https://lnkd.in/g_5fgAN8

10

Prometheus 2: open-source Language Model (LM) evaluation! Researchers have introduced Prometheus 2, a powerful open-source LM specifically designed for evaluating other LMs. This addresses growing concerns about the transparency, controllability, and affordability of relying solely on proprietary LMs like GPT-4 for evaluation purposes. Challenges with Existing Open-Source Evaluator LMs: *  Discrepancies with Human Judgement: Scores generated by these models often deviate significantly from human evaluations. *  Limited Evaluation Formats: They lack the flexibility to handle both direct assessments and pairwise rankings, which are the two most common evaluation methods. *  Restricted Evaluation Criteria: Existing models focus on general attributes like helpfulness and harmlessness, without the ability to incorporate custom criteria. Prometheus 2: A Game Changer *  Alignment with Human Judgements: Prometheus 2 demonstrates a remarkable ability to closely mirror human evaluations, as well as those of GPT-4. *  Versatile Evaluation Formats: It effectively handles both direct assessments and pairwise rankings. *  Customizable Evaluation Criteria: Users can define specific criteria to guide the evaluation process, allowing for greater control and precision. Results: Across various benchmarks, Prometheus 2 achieves the highest correlation and agreement with human judgements and proprietary LMs among all tested open-source evaluator LMs. Open Access: The models, code, and data for Prometheus 2 are all publicly available, fostering transparency and collaboration within the research community. Link to access the resources: https://lnkd.in/gudkqVyv Read more about the "Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models" in this research paper: https://lnkd.in/giW-rmRW #LanguageModels #LLM #AI #NLP #opensource #MachineLearning #LargeModels

26

Understanding Conversation Summary Memory vs Conversation Summary Buffer Memory in LangChain For detailed sample code, check out the Google Colab notebook(https://lnkd.in/gRsarpNX) In the evolving field of conversational AI, memory mechanisms play a critical role in maintaining coherent and contextually relevant interactions. Two key approaches, Conversation Summary Memory and Conversation Summary Buffer Memory, offer distinct ways of handling conversational context. While Conversation Summary Memory focuses on summarizing entire dialogues, making it ideal for long conversations where historical context is essential, Conversation Summary Buffer Memory retains a rolling window of recent exchanges, suited for short, focused conversations. Understanding the differences between these memory types and their appropriate use cases can significantly improve chatbot performance. For a deeper dive into how these memory systems function and their practical applications using LangChain, you can explore the full pdf article below. #ConversationalAI #AI #LangChain #Chatbots #ArtificialIntelligence #MemoryMechanisms #MachineLearning #NLP #OpenAI #AIChatbot #Technology

92

ANPD: Exciting advancements in LLM efficiency Researchers have developed a novel approach called Adaptive N-gram Parallel Decoding (ANPD) to address the resource consumption and latency challenges faced by LLMs. ANPD introduces a two-stage process: 1. Rapid Drafting: An adaptive N-gram module generates multiple tokens simultaneously, significantly speeding up the initial generation process. 2. Verification Phase: The original LLM meticulously evaluates and confirms the proposed tokens, ensuring the output's accuracy and fidelity. This innovative method retains the LLM's original quality while achieving impressive speed improvements. In experiments, models like LLaMA witnessed up to a 3.67x speedup! Key benefits of ANPD: * Enhanced Efficiency: Reduces resource consumption and latency. * Lossless Approach: Maintains the original LLM's output quality. * Plug-and-Play Enhancement: No retraining or additional GPU memory needed. ANPD represents a significant step forward in optimizing LLM performance, paving the way for wider adoption and more efficient applications. Read more about the Adaptive N-gram Parallel Decoding (ANPD) in this research paper: https://lnkd.in/gNzt2TK3 #LanguageModels #NaturalLanguageProcessing #AIResearch #MachineLearning #DeepLearning #Interpretability #ConceptBasedExplanations #CAVs #ModelUnderstanding #LLM #AI #Reasoning #NLP #MachineLearning #SelfPlay #AdversarialLearning #LLMs #AI #MachineLearning #NaturalLanguageProcessing #SelfLearning #Reasoning #Planning #scalinglaws #research #reproducibility #Breakthrough #ANPD #LanguageModels #Efficiency #Innovation

30

1 Comment

Addresses fundamental challenge of integrating contextual domain knowledge with RAG. Essential for real life applications in AI.

15

Hello Everyone Your learning - #learnwithvanama ---- 🔍 Exploring Best Practices in Retrieval-Augmented Generation (RAG) 📄 I recently delved into the insightful paper "Searching for Best Practices in Retrieval-Augmented Generation" by Xiaohua Wang et al., which offers a comprehensive survey on RAG techniques. This innovative approach merges large language models with information retrieval, enhancing content accuracy and relevance by tapping into external sources. Key Findings: - Mitigating Hallucinations: RAG proves effective in reducing inaccuracies and improving response quality, especially in specialized fields. - Multimodal Enhancements: Techniques for multimodal retrieval enhance question-answering related to visual inputs and expedite the generation of multimodal content. - Challenges: - Implementation complexities and extended response times are notable hurdles. - Integration Potential: Combining RAG with reinforcement learning or few-shot learning could yield more robust AI systems. Best Practices: - Continuous Evaluation: Regularly assess outputs to identify areas for enhancement. - Transparency: Clearly explain how outputs are generated to build user trust. - Personalization: Use customer integration data to tailor AI features effectively. The paper serves as a significant step forward in refining retrieval-augmented generation. While challenges persist, the proposed strategies and best practices offer a solid foundation for optimizing RAG and advancing AI capabilities. Embracing these insights can drive the development of more knowledgeable and contextually-aware AI systems. Link : https://lnkd.in/e3YJfq4q #AI #MachineLearning #Research #DataScience #Innovation #TechTrends

22

My latest Medium post comparing SetFit, FastFit, and Semantic Router for Few-shot NLP Intent Classification has been published by Towards AI! This post covers: 1. Details of NLP Intent classification and why it is extremely important in the age of LLM chatbots 2. Algorithm details of SetFit, FastFit and Semantic Router 3. Detailed code experiments on a chatbot intent detection dataset 4. Conclusion on the best method as of today. As with most of my posts, it has lots of code you can directly use in your work! This post was jointly written by Shilpa and me, both of us humans 😅 🔗 Medium Friend Link (without paywall): https://lnkd.in/dVfAvjzr Follow me for more tips on building successful ML and LLM products! Medium: https://lnkd.in/g2jAJn5 X: https://lnkd.in/g_JbKEkM #generativeai #llm #nlp #artificialintelligence #mlops #llmops

78

9 Comments

One more session alert by Vidvatta for people interested in prompt engineering. Topic - Efficient Prompt Engineering and Tuning for Language Models! 🗓️ Date: 26th May 2024 🕒 Time: 11 AM IST ⏳ Duration: 1 hour 30 mins 🔗 Register Here https://lnkd.in/gurGZmeV ✨ Join us in this webinar where you will have a unique experience of doing hands on together with mentor in this session and get live feedbacks when you are stuck. Are you interested in Large Language Models (LLMs) and their potential? Curious about how to leverage advanced prompting techniques for various tasks? Our upcoming webinar is perfect for you! ✨ What to Expect: A brief overview of LLMs and the importance of prompting. Prompting and Task Complexity- Classification of tasks: Easy, Medium, Hard. Advanced Prompting Techniques: * Zero-Shot and Few-Shot Learning: Definitions and applications. * Task Decomposition: Simplifying complex tasks. * Chain and Tree of Thought: Enhancing reasoning capabilities. Interactive Q&A Session: Open floor for attendee questions and sharing experiences. Future trends in LLMs and prompt engineering. Don't miss this opportunity to delve into the intricacies of LLMs and enhance your skills. Register now and be part of the conversation shaping the future of prompt engineering! In case of any queries contact [email protected] #Webinar #LLM #PromptEngineering #MachineLearning #ArtificialIntelligence #FutureTech #LearningCommunity

13

1 Comment

What does it take to put an ML model in production? The last session of BuildML covered the production readiness checklist for ML System. Here are the most important ones based on my experience 𝗙𝗲𝗮𝘁𝘂𝗿𝗲 𝗮𝗻𝗱 𝗗𝗮𝘁𝗮 𝗧𝗲𝘀𝘁𝘀 Capture Feature Expectation in Schema → Calculate statistics from train data, and mix with domain knowledge to create “expectation” of feature values. Test again this to make sure features aren't 'drifting' All Feature must be beneficial → Use leave-one-out feature-validation to trim feature set regularly (once a quarter or once a year) New Feature Can be added quickly → to allow teams to test with new feature, assess impact and integrate in as little as 1–2 months. 𝗠𝗼𝗱𝗲𝗹 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 𝗧𝗲𝘀𝘁𝘀 Every model specification undergoes code review to validate learning occurs and is checked into repo Impact of model staleness is known → assess is model isn't updated, what is the impact in offline and online metrics Model quality is sufficient on all important slices → check performance on important subset of data Model has been tested of bias → check performance on protected subsets like gender, race etc. 𝗠𝗟 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 Model specification is unit tested → ensure code is working using single run of gradient descent, the model is learning by ensuing training loss is decreasing etc ML Pipeline is integration tested → run with sample subset of dataset to ensure training and inference runs as expected Model quality is validated before serving → performance on held out validation set to check for performance degradation Model allows to debug/observe the results of a simple sample → framework like TF Debugger allows this Model is canary tested → risk is mitigated by running old and new version concurrently and gradually increasing traffic is new model is behaving well Model can be rolled back 𝗠𝗼𝗻𝗶𝘁𝗼𝗿𝗶𝗻𝗴 𝘁𝗲𝘀𝘁 𝗳𝗼𝗿 𝗠𝗟 Dependency changes are notified → upstream data changes will be notified to ML team Model does not experience reduction in prediction quality → drift detection on predictions, and tracking predictive performance - alert if performance degrades beyond certain point. Check out the full paper from Google here - https://lnkd.in/gACZrf8z Follow Karun Thankachan for everything DS/ML 𝗟𝗶𝗸𝗲 for more such content. 𝗥𝗲𝗽𝗼𝘀𝘁 to grow your own network!

87

26 Comments

F1-Score vs Accuracy: Why F1-Score Should Be Your Focus. In machine learning, it's tempting to focus on ACCURACY—the percentage of correctly predicted instances. But what if I told you that accuracy isn't always the best indicator of your model's performance? In cases where the data is imbalanced (e.g., fraud detection, disease diagnosis), accuracy can be misleading. Imagine a scenario where only 5% of transactions are fraudulent—if your model predicts "non-fraudulent" 100% of the time, you’ll get 95% accuracy but miss all the critical fraud cases! When it comes to the F1-score: - Precision tells you how many of your positive predictions were correct. - Recall tells you how many of the actual positives you correctly identified. The F1-score balances both precision and recall, giving you a better sense of how well your model performs—especially when false positives and false negatives matter. If you're working on models where missing a positive instance could have serious consequences (like healthcare, security, or finance), the F1-score should be your go-to metric. Focus on the right metrics to make smarter models and better decisions. #MachineLearning #DataScience #AI #F1Score #ModelEvaluation #TechInsights #DataDriven #BrainK

10

Hello all, I and my friend Aditya Sahu have been working lately on Automatic Redaction / Replacement tool for PDFs (scanned PDF & text-readable) and for Images which uses large language models (LLMs) in the backend. Different Formats / templates of PDFs cause a major problem in the extraction process. *What it does: 1. It removes the editing panel for your redaction / replacement work and instead you can just explain what you need to redact / replace in normal keywords or plain sentences in a comma separated fashion. 2. It uses LLMs in the background - you can use your own privately deployed model like llama 3 70b fine-tuned version or you can also use models like GPT-4o, Claude etc. which are commercially available. It can be configured based on your requirement. 3. It uses OCR for Image / Scanned Image case, which can be configured to any Cloud Server OCRs as well as open-source OCRs. 4. Can ingest multiple images and PDFs and work sync. 5. You can run multiple iterations on the same file to make many small small changes that can be possible in the PDF. Note: The process is not fully 100% accurate but it is well in range of 85-90+ , the accuracy can vary based on the choice of LLM model with good context length and choice of OCR as well as the quality of image. [*imp] *PDFs are quite complex for handling the formatting issues, so this tool does have limitations on what length of words can be replaced because it might lead to formatting issues in the provided PDF. There would be HIL [human in the loop] for evaluation. You can go through the Demo once; it has 2 parts in it - one for scanned PDF and one for text readable pdf (built using Python) 😀. Right now, it is just a Streamlit demo but parallelly in the backend work is going on... Please DM to me Mohit Singh Pawar or Aditya Sahu for more details or clarification related to its availability. Audio credits: ElevenLabs free credits. I guess will have to upload part 1 here and part 2 in a separate post. part 2 : https://lnkd.in/gEi-DUzb #redact #llm #replace #pdf #genai #generativeai

56

5 Comments

Gerhann Pieters brought up this very interesting discussion on ECL modeling. IMO, provisions analysts should always worry about the final impact of parameters models on the ECL. This is basically how their models are proven feasible. But... there is a world of derived analysis that can rise from this - with respect to impacts in both the PnL and the business strategy. That's why major banks set provision steering committees to collect insights and different perspectives.

23

India's AI Opportunity: Building Applications on Top of Large Language Models 🇮🇳🚀 In an exclusive interview with The The Economic Times, Andrew Ng, a prominent figure in the AI world, shared his insights on India's potential in the AI landscape. Here are the key takeaways: • India's biggest AI opportunity lies in building applications on top of large language models (LLMs) and foundation models. • While training LLMs is capital-intensive, the real value lies in the application layer, which is more accessible to startups and enterprises. • India should focus on sectors where its economy is already strong, such as IT services, manufacturing, agriculture, and finance, to maintain its advantages through AI applications. • The generative AI moment has leveled the playing field for startups and enterprises, with India standing at the verge of opportunities. • Ng's AI Fund invests in startups building AI applications, with an average of one investment per month. They have also invested in an Indian healthcare AI startup, Jivi Health. • Ng believes that with the rise of AI, learning to code will become a significant personal productivity booster, similar to language literacy. #AI #MachineLearning #DeepLearning #IndiaTech #Startups #TechTrends

1

 Dear LinkedIn Family, We are looking for Senior Data Scientist for Pune Location. #DataScientist LLM Architecture: Good understanding of the architecture underlying large language models, such as Transformer-based models and their variants. Design and implement deep learning model architectures using PyTorch. #LLM #PyTorch #pytorch Language Model Training and Fine-Tuning: Experience in training large-scale language models from scratch, as well as fine-tuning pre-trained models on domain data. Data Preprocessing for NLP: Skilled in preprocessing textual data, including tokenization, stemming, lemmatization, and handling of different text encoding. Transfer Learning and Adaptation: Proficiency in applying transfer learning techniques to adapt existing LLMs to new languages, domains, or specific business needs. Data Annotation and Evaluation: Skills in designing and implementing data annotation strategies for training LLMs and evaluating their performance using appropriate metrics. Scalability and Deployment: Experience in scaling LLMs for production environments, ensuring efficiency and robustness in deployment. Model Training, Optimization, and Evaluation: Evaluate the performance of PyTorch models using appropriate metrics and techniques like cross-validation, holdout sets, or online evaluation. This encompasses the complete cycle of training, fine-tuning, and validating language models. You will be designing and adapting LLMs for use in virtual assistants, Information retrieval and extraction etc. Experimentation with Emerging Technologies and Methods: Actively exploring new technologies and methodologies in language model development, including experimental frameworks and software tools. LLM Alignment: Understanding of algorithms like DPO, PPO, KPO, RLHF and using it for guardrails. AI Data Retrieval: Data retrieval from unstructured data, extract key value pairs using techniques like donut, layoutLM, table transformers. Analyze data and build EDAs to identify data patterns Hands-on and strong understanding of concepts in Deep Learning and NLP Proficient in TensorFlow and similar libraries.   Required Qualifications 5+ years of hands-on experience in developing and deploying Large Language Models, and Machine learning and working with Pytorch. A thorough understanding of machine learning, particularly deep learning techniques, including knowledge of neural network architectures, training methods, and optimization algorithms. Proficiency in AI technology, Python, including experience with NLP libraries (e.g., Hugging Face Transformers, NLTK, spaCy), text classification. Experience with frameworks: PyTorch, or Tensorflow. Experience with cloud services (AWS, Azure) and ML deployment tool Docker Familiarity with model fine-tuning and optimization techniques for LLMs. Proven track record of innovative solutions in the field of LLMs. Strong communication skills, with the ability to explain complex AI concepts to non-expert audiences.

4

1 Comment

Another great talk from the PyData London 2024 - Backtesting and error metrics for modern time series forecasting by Kishan Manani Backtesting is the time series forecasting equivalent to the machine learning cross-validation approach for training and testing models. https://lnkd.in/g-DWgqmB #timeseries #forecasting #machinelearning #python

46

5 Comments

A big shoutout to Kunal Raj for providing me with a sample PDF version of the "Mastering Transformers" book for review. It looks promising, covering many important topics from BERT to Stable Diffusion. I'll share my full review once I finish reading it! You can buy it here - https://packt.link/zfTHS

22

2 Comments

🚨Paper Alert 🚨 ➡️Paper Title: SGP-Bench: Can Large Language Models Understand Symbolic Graphics Programs? 🌟Few pointers from the paper 🎯Assessing the capabilities of large language models (LLMs) is often challenging, in part, because it is hard to find tasks to which they have not been exposed during training. Authors take one step to address this challenge by turning to a new task: focusing on symbolic graphics programs, which are a popular representation for graphics content that procedurally generates visual data. 🎯LLMs have shown exciting promise towards program synthesis, but do they understand symbolic graphics programs? Unlike conventional programs, symbolic graphics programs can be translated to graphics content. Here, Authors characterize an LLM's understanding of symbolic programs in terms of their ability to answer questions related to the graphics content. 🎯 This task is challenging as the questions are difficult to answer from the symbolic programs alone -- yet, they would be easy to answer from the corresponding graphics content as they verified through a human experiment. To understand symbolic programs, LLMs may need to possess the ability to imagine how the corresponding graphics content would look without directly accessing the rendered visual content. 🎯They used this task to evaluate LLMs by creating a large benchmark for the semantic understanding of symbolic graphics programs. This benchmark is built via program-graphics correspondence, hence requiring minimal human efforts. They evaluated current LLMs on their benchmark to elucidate a preliminary assessment of their ability to reason about visual scenes from programs. 🎯They found that this task distinguishes existing LLMs and models considered good at reasoning perform better. 🎯Lastly, they have introduced “Symbolic Instruction Tuning (SIT)” to improve this ability. Specifically, they queried GPT4-o with questions and images generated by symbolic programs. Such data are then used to finetune an LLM. They also found that SIT data can improve the general instruction following ability of LLMs. 🏢Organization: Max Planck Institute for Intelligent Systems, University of Cambridge, Mila - Quebec Artificial Intelligence Institute, Université de Montréal, University of Tübingen, Massachusetts Institute of Technology 🧙Paper Authors: Zeju Qiu, Weiyang Liu, Haven (Haiwen) Feng, Zhen Liu , Tim Z. Xiao, Katherine M. Collins, Joshua Tenenbaum, Adrian Weller, Michael J. Black, Bernhard Schölkopf 1️⃣Read the Full Paper here: https://lnkd.in/gu7E4VWX 2️⃣Project Page: https://lnkd.in/gud2vcGD 3️⃣Code: https://lnkd.in/geF-tvYR 🎥 Be sure to watch the attached 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.

8

📚 Book Review: "Mastering PyTorch" – Create and deploy deep learning models from CNNs to multimodal models, LLMs, and beyond by Ashish Ranjan Jha 🚀 I recently finished reading "Mastering PyTorch," and I must say it is an excellent resource for anyone looking to delve deep into the world of deep learning using PyTorch. Whether you are a data scientist, machine learning researcher, or deep learning practitioner, this book offers a wealth of knowledge and practical insights to enhance your skills and understanding. 🔑 Key Highlights: 1️⃣ Comprehensive Overview: The book kicks off with an introduction to deep learning and PyTorch, setting a strong foundation for beginners and experienced practitioners alike. 2️⃣ Advanced CNN and RNN Architectures: It covers SOTA advancements in Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), providing hands-on examples and practical applications. 3️⃣ Transformers and Hybrid Models: The exploration of transformers and hybrid models is particularly insightful, showcasing how these models have revolutionized natural language processing. 4️⃣ Graph Neural Networks (GNNs): A detailed dive into GNNs using PyTorch Geometric, explaining the concepts and implementations of Graph Convolutional Networks (GCNs) and Graph Attention Networks (GATs). 5️⃣ Generative Models: The sections on music and text generation, neural style transfer, GANs, and diffusion models are fascinating, demonstrating the creative potential of deep learning. 6️⃣ Deep Reinforcement Learning: Practical examples of training reinforcement learning agents in PyTorch add another dimension to the book's versatility. 7️⃣ Model Deployment: Step-by-step guidance on deploying PyTorch models into production, including mobile deployment for Android and iOS, is incredibly valuable for real-world applications. 8️⃣ Optimization and AutoML: Chapters on model training optimizations, distributed training, and AutoML provide techniques to efficiently train large models and automate the ML workflow. 9️⃣ Explainable AI: The focus on interpretability using tools like Captum ensures that the models are not just powerful but also understandable. 🔟 Integration with Hugging Face: Leveraging Hugging Face libraries with PyTorch to build cutting-edge AI models is a perfect ending to this comprehensive guide. If you have hands-on experience with Python and a basic understanding of PyTorch, this book is for you. It caters to those who wish to implement advanced deep learning paradigms and need practical examples to bridge the gap between theory and practice. 👨‍⚖️ "Mastering PyTorch" is more than just a technical guide; it is a gateway to mastering deep learning with one of the most powerful frameworks available today. 🌟 As an optim fan, Chapters 12, 13, 15 and 16 specially recommended! #DeepLearning #PyTorch #MachineLearning #AI #DataScience #BookReview #Transformers #ReinforcementLearning

26

5 Comments