Rajeev Ratan

🚀 𝗚𝗼𝗼𝗴𝗹𝗲 𝗗𝗲𝗲𝗽𝗠𝗶𝗻𝗱'𝘀 𝗔𝗜 𝗪𝗲𝗮𝘁𝗵𝗲𝗿 𝗙𝗼𝗿𝗲𝗰𝗮𝘀𝘁𝗶𝗻𝗴 𝗧𝗲𝗰𝗵 𝗪𝗶𝗻𝘀 𝗧𝗼𝗽 𝗨𝗞 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗔𝘄𝗮𝗿𝗱 Google DeepMind’s AI weather forecasting system, GraphCast, has won the prestigious 2024 MacRobert Award, marking a significant breakthrough in weather prediction. 🔑 Key Highlights: 1️⃣ 𝗥𝗲𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝗮𝗿𝘆 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆: GraphCast leverages advanced machine learning and vast datasets to deliver highly accurate and timely weather forecasts. 2️⃣ 𝗟𝗶𝗳𝗲-𝗦𝗮𝘃𝗶𝗻𝗴 𝗜𝗺𝗽𝗮𝗰𝘁: The system enhances the ability to predict severe weather events, enabling early safety warnings and better resource allocation, potentially saving lives. 3️⃣ 𝗔𝘄𝗮𝗿𝗱 𝗥𝗲𝗰𝗼𝗴𝗻𝗶𝘁𝗶𝗼𝗻: The MacRobert Award, presented by the Royal Academy of Engineering, honors outstanding engineering ingenuity, commercial success, and social benefit. 4️⃣ 𝗘𝘅𝗽𝗲𝗿𝘁 𝗘𝗻𝗱𝗼𝗿𝘀𝗲𝗺𝗲𝗻𝘁: GraphCast is praised for significantly improving the speed and precision of weather forecasting, crucial in the era of climate change. GraphCast showcases how engineering innovation can drive economic growth and foster a sustainable future. Precise weather forecasting is more critical than ever in addressing ongoing environmental challenges. 🌍🔬 #Innovation #Forecasting #AI #ArtificialIntelligence

3

I love creating parameterised plot functions. Yesterday, I created one to generate certificates for everyone on the latest dataviz workshop series after the course organisers asked if I could do that. Here's the workflow: - Download the group contact list from my emails - Read it into R - Send all the names through the plotting function with a quick one liner - Sip coffee while the pdfs are generated - Email every student their own certificate pdf One plot function, with one parameter: make_certificates(name), which makes use of ggplot2, ggtext, and ggimage to pop my logo in the background, and ggsaves with the student's name in the file name. I know not everything needs to be done in R, but it does make many things so much easier! I'm talking about creating parameterised plots at RMedicine next week. If you want to find out more about the process before then, I wrote a blog post about it a few weeks ago. https://lnkd.in/e_8nh7RJ

75

2 Comments

Following on from last week's article, this week I cover Activation Functions

1

Week 4 of DataTalksClub LLM-zoomcamp was rich in content. This week the focus was: 1) Setting up an LLM powered Streamlit app. The app provides users the capability to ask questions. We capture feedback from the user as part of the monitoring workflow. All data is stored and in a Postgres database. A visualisation dashboard is used for monitoring model performance (Grafana Labs). 2) We also explored how to use cosine similarity to deliver a verdict on how well an LLM is responding to user queries. A large language model was then used as a secondary evaluation method to guage quality of our user facing model. As always a big thanks to Alexey Grigorev and the team for making these available to many of us who enjoy collaborative learning and hands on projects. https://lnkd.in/drN-v7xv

12

Recently, I got to experiment on developing a Text vectorisation service for a continuous inflow of Tickets text data. We used Celery and E5 model to test the performance and tried to optimise on throughput. I have penned down my learnings and the issues I faced during the experiment Text vectorisation service using Celery and E5 model

21

As promised, this is the link to my article covering exhaustive details around building a RAG pipeline with bare models from sentence transformers. It also covers details around improving the retrieval efficiency by employing a re-ranking approach using a cross encoder model. Would love to hear your thoughts on this.! Find the link to my previous post on this in the comments section for more context. https://lnkd.in/gszrqPwA

19

11 Comments

Discover how eBay created their own language model using three billion item titles in this exciting video. We'll explore the innovative techniques eBay used, including training their BERT model from scratch on eBay's massive dataset, knowledge distillation to compress the model, and fine-tuning the compressed model to learn similarity better. With a 3.5% increased purchase order rate and enhanced customer engagement, this is a must-watch for anyone interested in the power of language models. Like and subscribe for more such interesting concepts. Also, like and share over here for maximum reach. : ) Video Link: youtu.be/h51nbWr7feo YT channel Link:  youtube.com/@datatrek #datatrek #datascience #machinelearning #statistics #deeplearning #ai

9

Tomorrow we have third and final session of our three session series on "Building Information Extraction from Text Books Using Gemma”. So far we have discussed about fundamentals of LLMs, Gemini and Gemma APIs, and building basic Q&A model using Gemma. Tomorrow we will discuss RAG-based application using Gemma. The event is organised by the TensorFlow User Group (TFUG) Jaipur in collaboration with Galgotias University IEEE Student Branch (IEEE GU). Hope to see you tomorrow!! 📅 Date: August 4 🕑 Time: 2 PM 🔗 Scan the QR code on the poster to register and for getting the event link. #TensorFlow #Gemma #Gemini #Google #MachineLearning #InformationExtraction #TFUGJaipur #IEEE #AI #DataScience #ProfessionalDevelopment

59

1 Comment

This week I have been thinking about security models with generative AI. Have a read and let me know your thoughts in the comments section! #MethodsForGood #ai #generativeai #security #cybersecurity #responsibleai# #dataprotection

9

🚀 Building AI with Foresight, Not Hindsight: The EPSRC Framework 🚀 The UK's Engineering and Physical Sciences Research Council (EPSRC) is pioneering a proactive approach to responsible AI with its AREA framework. Forget reacting to ethical dilemmas after they arise – AREA encourages us to anticipate the impact of AI before it's unleashed! 💡 Why AREA Matters: 💡 - Proactive Ethics: Tackling ethical considerations head-on, ensuring responsible AI development from day one. - Taming Autonomy: Guiding the development of autonomous systems, which are rapidly changing our world. - Practical Toolkit: Equipping developers and policymakers with the tools to navigate the ethical landscape of AI. The EPSRC framework is shaping a future where AI is not just innovative, but also ethical and beneficial for all. Let's build AI with foresight, not hindsight! #responsibleAI #AIethics #innovation #EPSRC #AREAframework #futureoftech

8

Really happy to share that our work on “Learning Metrics that Maximise Power for Accelerated A/B-tests” will appear in the Applied Data Science track of #KDD2024. We address a common problem in industry: how can we minimise type-II errors in online controlled experiments? Can we simply learn a metric combination that directly minimises this quantity? This project started out as “let’s implement the ideas from this great paper by Eugene Kharitonov et al.”, but we ended up proposing some extensions to their seminal work that seem to be quite effective, and are now used across ShareChat. Proud to have this work out there, let us know what you think. Additionally, we were assigned an Area Chair that clearly cared: initial low-confidence reviews prompted them to invite more reviewers (we got 8!), and the AC themselves correctly grilled our rebuttals to point out arguments they deemed insufficient. It’s easy to be positive about peer review when your paper gets accepted, but it seems there are still some unsung heroes out there... Work with Aleksei U.. ACM SIGKDD & Annual KDD Conference

116

9 Comments

Day 140 of the #365GenAIChallenge! 📅 Mmm pears... I mean today we have Byte Pair Encoding! Byte Pair Encoding is a data compression technique that iteratively replaces the most frequent pair of bytes in a sequence with a single, unused byte, effectively reducing the size of the data. In NLP, BPE is used to split words into subword units to handle out-of-vocabulary words and improve model performance.

3

Happy to share that I have just completed my AI Agents in LangGraph course! Multi-agent systems are critical for achieving non-trivial AI workflows, and utilising this technology can lead to significantly improved results in practice. Check out my accomplishment here: https://lnkd.in/gXzpyrD3

25

1 Comment

Gemma and Groq API: Enhancing AI and ML Workflows In the evolving world of AI and machine learning (ML), efficient tools are vital. Two significant APIs, Gemma and Groq, are designed to streamline model deployment and enhance computational performance. Gemma API Gemma simplifies AI and ML model deployment, offering: Ease of Use: Intuitive interface for quick deployment. Scalability: Handles varying loads, ideal for both small and large applications. Integration: Compatible with TensorFlow, PyTorch, and Scikit-learn. Security: Ensures secure data handling and compliance. Monitoring: Robust tools for tracking and managing performance. Groq API Groq is tailored for high-performance AI computations with: High Performance: Maximizes computational efficiency for deep learning. Custom Hardware: Optimized for Groq’s tensor processors. Low Latency: Ideal for real-time applications. Energy Efficiency: High performance with low power consumption. Developer-Friendly: Comprehensive documentation and support.

7

Day 44: Unveiling Local Structures with Locally Linear Embedding (LLE) #100DaysOfML Challenge Today on the #100DaysOfML Challenge, we explore Locally Linear Embedding (LLE), a dimensionality reduction technique that tackles high-dimensional data by focusing on local neighbourhoods. Unlike PCA or Isomap, which target global structures, LLE excels at preserving the local geometry of the data. Imagine a dataset scattered on a bumpy surface - LLE captures how each data point relates to its immediate neighbours, rather than the entire landscape. Here's a simplified look at LLE's approach: Nearest Neighbors: Similar to Isomap, LLE identifies the k nearest neighbours for each data point. Local Reconstruction: It assumes each data point can be reconstructed as a linear combination of its neighbours. Optimizing Weights: LLE finds the optimal weights for each neighbour that minimize the reconstruction error, essentially capturing the most important local relationships. Low-Dimensional Embedding: Finally, LLE projects the data points into a lower-dimensional space while trying to maintain these reconstructed relationships. LLE proves valuable for tasks like: Visualizing high-dimensional data with localized patterns: Explore how data points cluster or interact within their neighbourhoods. Dimensionality reduction for machine learning: Prepare data for algorithms that work better in lower dimensions, while preserving local relationships important for tasks like classification. Feature extraction: Identify underlying features within the data based on the learned local reconstructions. I'm interested in applying LLE to datasets with potential local variations and structures. Let's discuss! Have you used LLE before? What other dimensionality reduction techniques are you curious about? #MachineLearning #DataScience #LLE #DimensionalityReduction #LocalGeometry #100DaysOfMLChallenge

3

Phi-3 Vision is now available on Huggingface. Phi series models from MS have smaller sizes but higher accuracy in their size range because the models are trained on a curated high quality dataset. "The model is intended for broad commercial and research use in English. The model provides uses for general purpose AI systems and applications with visual and text input capabilities which require: >> memory/compute constrained environments; >> latency bound scenarios; >> general image understanding; >> OCR; >> chart and table understanding. Our model is designed to accelerate research on efficient language and multimodal models, for use as a building block for generative AI powered features." Mithilesh Thakkar

9

Latest Development & Release In Generative AI LLM Space #latest #trending #development #released #opensource #llm #genai Gemma 2 a new lightweight Opensource model released by Google for developers and researchers. Available in both 9 billion (9B) and 27 billion (27B) parameter sizes, Gemma 2 is higher-performing and more efficient at inference than the first generation, with significant safety advancements built in. #gemma2 #genai #google #development #aiengineer #ai #meta #vertexai Gemma 2 model outperforming Llama3 (Meta AI)with several benchmark evaluation especially General Reasoning with 86.4 score. Know Details With Model Card : https://lnkd.in/gsgYSRfE #quantization #robotics #embedai #opensource #mac #kaggle Quantized Version : https://lnkd.in/gPy3jK4Q Key Architecture Points To Consider for Gemma 2: 1. Rotary Positional Embeddings 2. GeGLU 3. Grouped Query Attention with num_groups = 2 4. Context Length 8192 5. Full Precision fast inference on Single TPU , Nvidia H100 Full Technical Report : https://lnkd.in/gFiQpkJx Fine-Tuning Gemma2 With Keras : https://lnkd.in/gR7MZ_8U #nvidia #generativeai #gpt #openai #finetuning #development #artificialintelligence #generativedesign #architect #solutionarchitect #azure #huggingface #keras #llmops #mlops #product #vectordb #consulting #consultant #meta #AI

7