Shiwei(George) Huang

🌟 Excited to share my latest contribution to the AWS Machine Learning Blog! 🌟 In this post, we delve into how Amazon trains Sequential Ensemble Models at scale using Amazon SageMaker Pipelines. This approach has enabled us to: ✅ Streamline model orchestration, ✅ Scale training efficiently, and ✅ Enhance predictive accuracy for complex business problems. This work showcases the power of Amazon SageMaker in tackling large-scale machine learning workflows while ensuring scalability and efficiency. 🚀 Check out the blog to learn more: #MachineLearning #SageMaker #AWS #AI #DataScience

9

Participants will design and implement a basic chat agent using RAG techniques with NVIDIA-optimized inference software of NIM microservice. This challenge aims to provide hands-on experience with LLM systems, dialog management, and document retrieval with NIM acceleration.

2

👉 Amazon OpenSearch Ingestion adds support for ingesting data from self-managed sources Amazon OpenSearch Ingestion now allows you to ingest data from self-managed OpenSearch, Elasticsearch and Apache Kafka clusters, eliminating the need to run and manage 3rd party tools like Logstash to migrate your data from self-managed sources into Amazon OpenSearch Service. Now you can seamlessly migrate or continuously replicate your data from all OpenSearch versions and Elasticsearch 7.x versions either on Amazon EC2 or on-premises environments into Amazon OpenSearch Service managed clusters or Serverless collections. #aws #amazonopensearch #seamlessingestion

6

This research from Facebook efficiently hard-codes large amounts of information into model memory layers, without increasing the number of required computations. It matches the capability of models roughly 4x larger. It uses key-value pairs within the actual model parameters that can be recalled highly efficiently, this is especially important for fields where factual accuracy is important (100% improvement in factual recall on QA benchmarks). There was also some custom CUDA Kernels to overcome memory bandwidth bottlenecks for the EmbeddingBag operation (used to recall the most appropriate information based on the query), and other cool stuff. Facebook clearly has multiple independent teams converging with all this research, I think a combination of all these breakthroughs -rather than a single event- is really going to move us forward. https://lnkd.in/ecC7MW_U

4

1 Comment

🔓 Unlock the power of your data with NVIDIA NeMo Curator. Discover how to #customize and optimize your model by integrating high-quality, domain-specific datasets. https://nvda.ws/3WO1QYZ

1

check out how AWS SakeMaker use FSDP2 + float8 from TorchTitan and TorchAO for 38% speed up

8

Happy to share our latest work "MSc-SQL: Multi-Sample Critiquing Small Language Models For Text-To-SQL Translation" published in the Table Representation Workshop, NeurIPS 2024. Text-2-SQL translation is an important problem to enable agentic workflows for automating database tasks. Recent advances in the space of text-to-SQL utilize closed proprietary models like GPT-4 that present challenges in accessibility, privacy, and latency. We address these issues by developing small, efficient open models (under 10 billion parameters) for text-to-SQL translation and obtain state-of-the-art results compared to other open models while remaining competitive with larger proprietary models at a much lower cost. Joint work with my colleagues at Layer 6 AI. Ilan Gofman, Zhaoyan Liu, Paul Wu, Noël Vouitsis, Guangwei Yu, Jesse Cresswell, Rasa Hosseinzadeh. Paper link: https://lnkd.in/ghi5mKwB

98

1 Comment

🎯 How teams are cutting Machine Learning Engineer hiring time from 90+ days to 28 days: 1. Role-specific technical assessments that mirror real projects • Real-world problem-solving instead of whiteboard exercises • Actual code reviews from your codebase • System design challenges based on your architecture • Practical ML model optimization tasks 2. Direct access to tech leads in first interview • No HR screening delays • Technical discussion from day one • Immediate architecture alignment • Quick technical fit assessment 3. 48-hour feedback commitment from all parties • Same-day interview feedback • Next-steps within 24 hours • Offer decisions in 48 hours • Clear communication timeline 4. Streamlined technical evaluation process • One comprehensive technical round • Parallel evaluations when possible • Combined culture/technical assessments • Quick technical team consensus Teams implementing even one of these cut hiring time by 40%. "Especially effective for fast-growing startups and scale-ups" #TechHiring #MLEngineering #TechRecruitment

3

I’ve just published a new article on Medium: "Leveraging LangChain MapReduce: Summarize Large Texts and Translate into Any Language." Disclaimer: While this solution offers a fresh perspective on summarizing and translating large texts, it may not be foolproof. Consider this as an innovative approach to exploring the problem space, rather than a one-size-fits-all solution.

14

Earlier this week, we announced an update to how Amazon Q Developer Software Agent works - https://lnkd.in/gVmTUQhA. Our efforts since making Q Developer Generally Available in May '24 have propelled the agent for software development to new heights. The SWE-bench scores have increased significantly, rising from 25.6% to 38.8% on the verified dataset and from 13.82% to 19.75% on the full dataset. This progress is a result of our customer-centric approach, incorporating developer feedback, and rapidly iterating to address pain points. What sets our agent apart is its architecture, which not only leverages the best LLMs but also equips the agent with the ability to explore multiple paths continuously to find optimal solutions for problems, similar to how developers work. There is a FREE version of Amazon Q Developer available (https://lnkd.in/gJdU3pB5) for you to try and experiment with. Install or update to the latest version of the Q Developer IDE extension, then type "/dev" in the chat window to put the Q Developer Agent to work.

37

🚀🚀🚀 𝐍𝐞𝐰 𝐀𝐫𝐭𝐢𝐜𝐥𝐞 𝐀𝐥𝐞𝐫𝐭 🚀🚀🚀 𝐅𝐢𝐧𝐞-𝐓𝐮𝐧𝐢𝐧𝐠 𝐋𝐚𝐫𝐠𝐞 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐌𝐨𝐝𝐞𝐥𝐬 𝐰𝐢𝐭𝐡 .𝐍𝐄𝐓 𝐂𝐨𝐫𝐞, 𝐏𝐲𝐭𝐡𝐨𝐧, 𝐚𝐧𝐝 𝐀𝐳𝐮𝐫𝐞 Large Language Models are redefining how we approach natural language understanding, automation, and decision-making across industries. This visualization captures the interconnected intelligence and global impact of LLMs, symbolizing their role in shaping the future of technology. From powering AI-driven customer experiences to advancing research and innovation, LLMs are at the forefront of this transformation. 💡 How do you see LLMs impacting your industry? Let’s discuss! #dotnet #azure #python #LLM #AI https://bit.ly/4j60Sjk

36

19 Comments

Learn why the NVIDIA Jetson AGX Orin is the leading platform for running transformer models like GPT-J, vision transformers, and stable diffusion at the edge.

1

🔍 Exploring Sparse Distributed Memory (SDM) I'm excited to share my latest blog post on Sparse Distributed Memory (SDM), an innovative model inspired by human memory! 🚀 In this detailed walkthrough, I break down the SDM script I made, providing rich mathematical and scientific insights to help you understand and implement this robust and scalable memory storage and retrieval system. If you're curious about how SDM can serve as a fascinating alternative to traditional neural networks, enhancing your machine learning projects, check out the full article here. #MachineLearning #ArtificialIntelligence #SparseDistributedMemory #SDM #DataScience #NeuralNetworks #MemoryModels #TechInnovation #Coding #AIResearch

4

"Can LLMs introspect? We define introspection as acquiring knowledge that is not contained in or derived from training data but instead originates from internal states. Such a capability could enhance model interpretability. Instead of painstakingly analyzing a model's internal workings, we could simply ask the model about its beliefs, world models, and goals. More speculatively, an introspective model might self-report on whether it possesses certain internal states such as subjective feelings or desires and this could inform us about the moral status of these states. Such self-reports would not be entirely dictated by the model's training data." https://lnkd.in/e2Cs3WA2

3