Virtual Highlights From ODSC APAC 2024 Now Available On-Demand
ODSC APAC 2024 has come and gone and we couldn’t be more thankful to our wonderful speakers and attendees for sharing their knowledge, insights, and passion for AI and data. If you couldn’t join us for the conference, you can still check out the sessions below. You can watch them all on-demand here at any time!
Retrieval-Augmented Generation (RAG): A Synergistic Approach to Natural Language Understanding and Generation
Shalvi Mahajan | Senior Data Scientist | SAP SE
This talk dives deep into the Retrieval-Augmented Generation (RAG) paradigm, which represents a novel architecture at the intersection of retrieval and generation models, designed to address the challenges in natural language understanding and generation tasks. This approach seamlessly integrates the strengths of information retrieval techniques with the creative capabilities of natural language generation, thereby achieving enhanced performance across a spectrum of language-based applications.
Practical Takeaway for Language Models and LLM in Production
Kevin Noel | Lead of AI | Uzabase Speeda
This presentation delves into some of the practical takeaways of using LLM in a production application. You’ll start with general machine learning concepts, such as Representation Learning and multi-task learning, and continue with an overview of different tasks that LM and LLM can achieve, practical differences between LM and LLM, and the current architectures used for LM and LLM.
The presentation provides a practical overview of LMs and LLMs, highlighting the fundamentals and concrete use cases to be customized for specific applications.
Multi-channel Multi-granularity Forecasting on High-Velocity Data
Thejas Muralidhar Bhat | Senior Quantitative Analytics Specialist | Wells Fargo
Naveen Rathani | Data Science & Analytics Lead | Wells Fargo
This session dives deep into the efforts to build, test, and deploy hundreds of models that can provide micro-level forecasts for transactional and operation data to the business at various aggregation levels and across various times the granularities to facilitate both tactical and strategic decision-making. You’ll also explore the solution that the speakers developed and how they did so.
Graph Data-Driven Recommendation Systems Empowered by Generative AI
Ravi Ranjan | Principal Data Scientist | Publicis Sapient
In this session, you’ll explore the challenges of using recommendation systems for travel planning and the possible solutions offered by the advent of Generative AI. Discover how Generative AI, paired with innovative modeling, can transform a user’s journey from decision to destination, creating personalized, timely, and engaging recommendations.
Securing Mobile Transactions: Large Language Models and the Future of FinTech Apps
Sanjaikanth Pillai | Staff Site Reliability Engineer | Visa Inc.
In this session, you’ll delve into various Low Latency Verification (LLV) trends in FinTech on mobile handheld devices, exploring the applications of LLM in FinTech and the potential of LLM alongside AI/ML in FinTech. In addition, you’ll address the limitations of LLM in achieving full implementation strength.
Going From Unstructured Data to Vector Similarity Search
Steven Pousty, PhD | Principal and Founder | Tech Raven Consulting
One of the key concepts used in AI modeling is the storage and query of vectors. In this workshop, you will take 2 forms of unstructured data, images, and journal abstracts, and work them all the way through to a usable data store with an application on top of it.
You’ll cover things such as transformers, embeddings, HuggingFace, choosing a vectorization model, and effective query composition.
Ascribing Responsibility for Unintended Consequences of AI Influence
Dr. Seema Chokshi | CEO & Founder | DataWyz.ai
This talk focuses on the crucial issue of how to attribute responsibility when AI systems lead to unintended consequences by influencing human decision-making. You’ll be introduced to the concept of the “decision point dilemma,” where it becomes unclear whether the human or the AI should be held accountable for negative outcomes resulting from AI-swayed decisions.
In particular, you’ll explore the relevance and importance of addressing this dilemma, as the increasing ubiquity of AI in assisting human choices is likely to amplify unintended AI influence, potentially leading to biased decisions, loss of human autonomy, misuse of AI, and erosion of accountability.
Leveraging LLMs for Internal Knowledge Discovery
Raghav Bali | Staff Data Scientist | Delivery Hero
Timothy Low | Senior Data Scientist | foodpanda
Efficient knowledge discovery is crucial yet challenging in today’s fast-paced, globally distributed work environments. This talk introduces a Retrieval Augmented Generation (RAG) based bot integrated with internal communication tools such as Slack. It is designed to streamline access and discovery of documentation spread across multiple internal knowledge platforms code-repositories and multiple languages. During it, you will explore the chatbot’s innovative architecture and practical applications.
Composability in AI Systems
Hugo Bowne-Anderson, PhD | Independent Data and AI Consultant
In this talk, you’ll explore how combining different AI tools and models can lead to innovative applications. In particular, you’ll discuss a voice-to-image system that integrates speech-to-text conversion, text processing, and image generation, demonstrating the power of composability in AI.
Real-world examples from programming and data science, like the pipe operator in R and scikit-learn pipelines, will illustrate how composability works in practice. Additionally, you’ll discuss the importance of structured inputs and outputs and tools like Pydantic and LangChain that help manage these structures effectively.
Building Your First Multimodal Gen AI App
Hugo Bowne-Anderson, PhD | Independent Data and AI Consultant
Don’t miss this practical, hands-on workshop where you’ll learn to build your very own multimodal generative AI app. You’ll learn how to combine state-of-the-art models from platforms such as Replicate and Hugging Face, as well as large language models like Bard, Claude, Mistral, and ChatGPT.
Mastering AI Image Generation for Creatives: From Concept to Stunning Visuals
Dr. Seema Chokshi | CEO & Founder | DataWyz.ai
This hands-on tutorial will empower creative professionals and content creators to leverage AI image generation tools like Midjourney and Stable Diffusion to rapidly ideate, prototype, and produce stunning visuals. Through guided exercises and expert tips, participants will learn to craft effective text prompts, iterate on AI-generated images, and integrate them into professional design projects. By the end of the session, attendees will have gained practical skills to start using AI to amplify their creative output and streamline their workflows.
Operationalizing GenAI: Effective LLM Compression and Optimization Methods
Nijesh Kanjinghat | AI Engineering Lead | IBM
This session delves into cutting-edge techniques to streamline the deployment of generative AI applications. It is designed for AI practitioners, data scientists, and machine learning engineers who are keen on enhancing their expertise in model optimization and seeking practical solutions for real-world applications. You’ll explore a range of LLM compression techniques, including quantization, pruning, and knowledge distillation, and understand how these methods can significantly reduce model size and inference costs without sacrificing performance. The session will also cover optimization strategies that improve model efficiency and speed, making it easier to deploy LLMs across various platforms and environments.
What’s next?
You might have missed ODSC APAC, but you can still join us at ODSC Europe. So what are you waiting for? Get your pass today before they run out!
Originally posted on OpenDataScience.com
Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday. You can also get data science training on-demand wherever you are with our Ai+ Training platform. Interested in attending an ODSC event? Learn more about our upcoming events here.
