Guillermo Breto Rangel PhD.

I am planning to post couple of interesting and award winning papers from #NeurIPS for next coupe of days:  Title: 𝗡𝗼𝘁 𝗔𝗹𝗹 𝗧𝗼𝗸𝗲𝗻𝘀 𝗔𝗿𝗲 𝗪𝗵𝗮𝘁 𝗬𝗼𝘂 𝗡𝗲𝗲𝗱 Selective Language Modeling for Efficient LLM Training Paper Link: https://lnkd.in/eXFBGDAU AI Generated Podcast: https://lnkd.in/eXH7CuX8 In the evolving landscape of AI research, this paper on Selective Language Modeling (SLM) offers intriguing perspectives on training efficiency. The RHO-1 approach introduces a nuanced method of language model training. By using a reference model to score token importance, the research challenges traditional data-intensive training paradigms. 𝗞𝗲𝘆 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗢𝗯𝘀𝗲𝗿𝘃𝗮𝘁𝗶𝗼𝗻𝘀: • Intelligent Token Selection: Researchers explored a method of focusing model training on tokens with higher informational utility • Data Efficiency Approach: Demonstrated potential for improved performance using fewer training tokens • Precision in Machine Learning: Highlighting the value of targeted training strategies 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝗮𝗹 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀: • Improvements in few-shot learning capabilities • Enhanced accuracy across mathematical and general task domains • Potential for more resource-efficient model development #MachineLearning #AIResearch #TechInnovation #NeurIPS

15

### 𝗠𝗲𝗺𝗼𝗥𝗔𝗚: 𝗔 𝗥𝗲𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝗮𝗿𝘆 𝗥𝗔𝗚 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 𝗳𝗼𝗿 𝗘𝗻𝗵𝗮𝗻𝗰𝗲𝗱 𝗘𝘃𝗶𝗱𝗲𝗻𝗰𝗲 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹! MemoRAG is transforming the way we approach retrieval-augmented generation (RAG) by incorporating a super-long memory model for global understanding across massive datasets. 🌐 In real-world applications, memory plays a critical role. For instance, in a knowledge-intensive question-answering task, using MemoRAG improved response accuracy by 27% compared to traditional RAG frameworks. Such gains highlight the value of global memory in providing nuanced, contextually rich answers. #### 🔥 𝗞𝗲𝘆 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀 🔥 1. 𝗚𝗹𝗼𝗯𝗮𝗹 𝗠𝗲𝗺𝗼𝗿𝘆: Processes up to 1 million tokens in a single context, offering unmatched depth and breadth across data. 2. 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝗯𝗹𝗲 & 𝗙𝗹𝗲𝘅𝗶𝗯𝗹𝗲: Fine-tune and adapt MemoRAG to new tasks with only a few hours of training! 3. 𝗖𝗼𝗻𝘁𝗲𝘅𝘁𝘂𝗮𝗹 𝗖𝗹𝘂𝗲𝘀: Bridges raw inputs to answers using clues derived from memory, unlocking insights from even the most complex queries. 4. 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗖𝗮𝗰𝗵𝗶𝗻𝗴: Up to 30x faster context pre-filling through advanced caching, chunking, indexing, and encoding. 5. 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗥𝗲𝘂𝘀𝗲: Long contexts can be encoded once and reused, improving efficiency for tasks with repetitive data needs. 🔄 𝗟𝗶𝘁𝗲 𝗠𝗼𝗱𝗲 𝗼𝗳 𝗠𝗲𝗺𝗼𝗥𝗔𝗚 Experience MemoRAG’s powerful pipeline with just a few lines of code! Ideal for GPUs with 16GiB or 24GiB memory, the Lite Mode simplifies getting started while maintaining exceptional performance. #### ✨ 𝗕𝗮𝘀𝗶𝗰 𝗨𝘀𝗮𝗴𝗲 & 𝗠𝗼𝗱𝗲𝗹 𝗖𝗼𝗺𝗽𝗮𝘁𝗶𝗯𝗶𝗹𝗶𝘁𝘆 MemoRAG works seamlessly with HuggingFace models, using the `MemoRAG.memorize()` method to build global memory across long input contexts. #### 🧠 Long LLMs as Memory Models MemoRAG supports long-context LLMs like Meta-Llama-3.1-8B-Instruct and Llama3.1-8B-Chinese-Chat, optimizing memory through MInference. #MemoRAG #RAGFramework #LLMs #GlobalMemory #AI #NLP #MInference #GenerativeAI #MasteringLLM

88

2 Comments

𝗨𝗻𝗹𝗼𝗰𝗸𝗶𝗻𝗴 𝗔𝗜 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆: 𝗛𝗼𝘄 𝗠𝗶𝗻𝗶𝘁𝗿𝗼𝗻 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝘀 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗶𝗻𝘁𝗼 𝗟𝗲𝗮𝗻 𝗣𝗼𝘄𝗲𝗿𝗵𝗼𝘂𝘀𝗲𝘀 (Paper link: https://lnkd.in/eJX8emrQ AI Generated Podcast: https://lnkd.in/eRUd-crt) NVIDIA researchers have developed the Minitron approach to compress large language models (LLMs) like Llama 3.1 8B and Mistral NeMo 12B into smaller, more efficient versions with 4B and 8B parameters, respectively. This is achieved through 𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲𝗱 𝗽𝗿𝘂𝗻𝗶𝗻𝗴 and 𝗸𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗱𝗶𝘀𝘁𝗶𝗹𝗹𝗮𝘁𝗶𝗼𝗻, enabling significant reductions in model size without substantial performance loss. 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀: • Pruning Strategies: Two methods were explored: depth pruning and joint hidden/attention/MLP (width) pruning, effectively reducing model complexity. • Teacher Correction Phase: Introducing a fine-tuning step allows the teacher model to adapt to specific data distributions, enhancing the distillation process. • Open-Source Models: The resulting models, including Mistral-NeMo-Minitron-8B, are available on Hugging Face with permissive licenses, promoting accessibility and further research. This advancement offers a practical pathway to deploy powerful language models in resource-constrained environments, broadening their applicability across various industries. #ArtificialIntelligence #MachineLearning #ModelCompression #AIResearch #OpenSourceAI #EfficientAI #LLM #TechInnovation

9

1 Comment

A very useful checklist to follow when you have to put in production Transformer based models

5

𝗖𝗮𝗻 𝗔𝗜 𝗧𝗲𝗹𝗹 𝗦𝘁𝗼𝗿𝗶𝗲𝘀 𝗟𝗶𝗸𝗲 𝗛𝘂𝗺𝗮𝗻𝘀? 𝗘𝗻𝘁𝗲𝗿 𝗠𝗶𝘅𝘁𝘂𝗿𝗲-𝗼𝗳-𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝘀 (𝗠𝗼𝗣) A new strategy called 𝗠𝗶𝘅𝘁𝘂𝗿𝗲-𝗼𝗳-𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝘀 (MoP) is redefining how Large Language Models (LLMs) evaluate stories. By prompting LLMs to adopt personas focused on specific storytelling metrics—like empathy, authenticity, or narrative complexity—MoP improves alignment with human judgment by over 20%. Psychological Depth Scale (PDS): A New Lens for Stories Using the PDS framework, researchers assessed the psychological depth of LLM-generated short stories based on: • Authenticity: Genuine human-like emotions. • Empathy: Evoking connection with characters. • Engagement: Captivating readers. • Emotion Provocation: Stirring strong reactions. • Narrative Complexity: Intricate plots and characters. 𝗞𝗲𝘆 𝗙𝗶𝗻𝗱𝗶𝗻𝗴𝘀 • LLMs rival humans: Models like GPT-4 created stories with depth comparable to human-written ones. • Size isn’t everything: Smaller models (e.g., Llama-2-7B) performed surprisingly well. • Emotional AI: MoP showed that LLMs could reflect diverse, nuanced human perspectives. 𝗪𝗵𝘆 𝗜𝘁 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 This research shifts the focus from technical quality to emotional and intellectual engagement, opening doors for AI in education, entertainment, and even therapy. Can storytelling AI become the next creative frontier? What do you think? Paper Link: https://lnkd.in/ewq23aFA Spotify Episode of this paper: https://lnkd.in/enqEe5TZ #AI #Storytelling #GenerativeAI #PsychologicalDepth #MixtureOfPersonas #LLM #NotebookLM #Innovation

5

👀 Discover how GPU acceleration is transforming causal inference. Using RAPIDS cuML with DoubleML, you can tackle large datasets up to 12x faster. ➡️ https://nvda.ws/3YRM4eQ

161

1 Comment

Self-Correcting Code Generation A new project demonstrates the power of Codestral for Self-Correcting Code Generation applications. Codestral has introduced an innovative self-correcting code generation approach by combining the capabilities of its generative model with AlphaCodium's self-correction methodology. This implementation leverages LangGraph for structured code output and feedback, making significant strides in automating and refining code generation processes. For developers and tech enthusiasts looking to reduce debugging time and improve code quality, this tutorial offers valuable insights into the next generation of coding assistants. Explore the detailed steps and examples to harness the power of self-correcting code generation. #CodeGeneration #AI #MachineLearning #Coding #Technology #SoftwareDevelopment #DeveloperTools #Automation #Innovation #DeepLearning #NLP #NaturalLanguageProcessing

6

Let me present you the 2nd edition of "Handbook of Anomaly Detection" in response to the demand for the 1st edition. This book is special to me on its dedication page to Eileen Kuo: "To our beloved daughter, Eileen, You are the blossom that brightens us with every smile, and fills our lives with beauty, joy and love. Mom and Dad" I write with care and love, and I believe you will read with joy and smile. The second edition is an expanded edition including: - Comprehensive and in-depth coverage - Enhanced explanations and visual presentations - Supervised and unsupervised learning techniques - 200+ data science Q&A Check it out! Amazon print book: $65 https://a.co/d/c9Yeqph The print edition adopts glossy cover, color print, and the beautiful Springer font and layout for pleasing reading. Its 7.5 x 9.25 inches portal size goes with most of your books in your bookshelf. Sample eBook chapters: free https://lnkd.in/eG8HSChZ eBook on Teachable: $22.50 https://lnkd.in/eeMmwpFz This eBook is in a beautiful format for a pleasing reading experience. We have chosen the Teachable platform for the distribution of the eBook without hefty overhead and the distribution to readers around the world. The credit card transactions are managed confidentially and securely.

17

2 Comments

Join us as we sit down with Ernest Chan—physicist-turned-quant, author, and founder of PredictNow.ai. In this episode, Dr. Chan shares invaluable insights on building robust data pipelines, engineering predictive signals, and choosing the right ML models for live trading. We also dive into practical risk management techniques, dynamic position sizing, and the role of AI agents in asset management. https://lnkd.in/gSikSjxc 

55

4 Comments

LLM interview question: Unlocking the Power of Prompt Chaining: A Guide to Mastering Generative AI 🚀 In the ever-evolving world of generative AI, prompt chaining has emerged as a powerful technique for unlocking the full potential of large language models like me 🤖. But what exactly is it, and how can you leverage it for your own benefit? 🤔 What is Prompt Chaining? 🔗 Prompt chaining involves feeding the AI model a series of interconnected prompts, each building upon the previous one. This allows the model to maintain context and understanding throughout the interaction, leading to more nuanced and relevant outputs 💡. When to Use Prompt Chaining: ⏰ Here are some key scenarios where prompt chaining shines: * Complex Tasks: Breaking down intricate tasks into smaller, sequential prompts guides the model towards a specific goal, ensuring accuracy and efficiency 🎯. * Creative Content Generation: Chaining prompts allows for the gradual development of ideas, leading to more elaborate and cohesive creative outputs, like poems, scripts, or musical pieces 🎨🎶. * Iterative Refinement: By providing feedback through subsequent prompts, you can continuously refine the model's response, tailoring it to your precise needs 🔄. Why Use Prompt Chaining? 🤔 The benefits of prompt chaining are numerous: * Enhanced Coherence: By maintaining context across prompts, the model generates outputs that are more internally consistent and logically sound 🧠. * Improved Accuracy: Breaking down tasks into smaller steps reduces the risk of errors and ensures the model stays on track 💯. * Greater Control: Chaining allows you to guide the model's thought process, leading to outputs that align perfectly with your vision 🧭. How to Use Prompt Chaining Effectively: 💪 Here are some tips for maximizing the effectiveness of prompt chaining: * Start Simple: Begin with a clear and concise initial prompt, establishing the foundation for the subsequent chain 🧱. * Logical Progression: Ensure each prompt builds upon the previous one, maintaining a clear flow of information and reasoning ➡️. * Provide Context: Where necessary, provide additional context within the prompts to guide the model's understanding 💡. * Refine and Iterate: Don't be afraid to experiment with different prompt sequences and provide feedback to achieve the desired outcome 🔁. By mastering the art of prompt chaining, you can unlock the full potential of generative AI, pushing the boundaries of creative expression, problem-solving, and more 🚀. Let's leverage this powerful technique together! Share your thoughts and experiences with prompt chaining in the comments below 👇. Follow Rohan Chikorde for more such AI content. #generativeai #promptchaining #ai #machinelearning #nlp #creativity #problem-solving #aipowered #futurism #techinnovation #llms #gpt3 #bard #openai #machinelearningengineer #datascience #bigdata #coding #programming #futureofwork #automation #deeplearning

16

2 Comments

Interesting insights. I think the MLFlow project did a good thing by providing far more information on different deployment patterns in their docs. Has helped move people from “I’ll run locally and have a nice UI but create an un-manageable mess” to “even if I don’t buy something, with a bit of work I can do something more centralised and sensible”. I think this will have helped their (already high) adoption, especially now in the face of stiff commercial vendor competition. Fun times for #mlops

21

Check out our latest video where we break down IndoxGen's key data generation pipelines: SyntheticDataGenerator: LLM-powered with quality control. SyntheticDataGeneratorHF: Advanced pipeline with human feedback integration. Watch now to see how these pipelines can supercharge your synthetic data projects! 🔗 Learn more: https://lnkd.in/de7PpuuE #AI #SyntheticData #MachineLearning #IndoxGen #LLM

7

🔥 LLMOps. LLMEval with #Opik library from Comet. #python #LLMs #llmops #ml #llmeval

5

In this episode of Signals & Threads, Sylvain and Ron go deep on how to make training more efficient, the subtle performance bugs #PyTorch lets you write, how to keep a hungry GPU well-fed, and the importance of reproducibility in training runs. They also discuss some of the unique challenges of doing #machinelearning in the world of trading, like the unusual size and shape of market data and the need to do inference at very low latencies. https://lnkd.in/emnkBYRH

573

5 Comments

In context of debate if large context #LLM models will eventually negate the need for #RAG, an article from Databricks has some good insights - • Longer contexts boost RAG accuracy... to a point • Performance often peaks at 32k-64k tokens, then declines • Top models (GPT-4, Claude 3.5) maintain consistency at scale • Others show unique failure modes (e.g., copyright concerns, summarization instead of Q&A) • Optimal context size varies by model and task • Lack of long-context training data may explain some issues Key takeaway: Long context and RAG are synergistic, but LLMs need refinement to fully leverage extended contexts. https://lnkd.in/eehdKtCm

17

🚀 Check out the latest Medium post by our student Tianyunxi Yin: "Quantization Fundamentals for Model Compression". Explore the essentials of quantization and learn how it helps in compressing machine learning models for better efficiency without compromising performance. Read more here: https://lnkd.in/gHbbR_J4 #USFMDS #DataScience #MachineLearning #ModelCompression #Quantization

15