The Data Exchange with Ben Lorica

Knowledge graph experts from data.world discuss their work on using knowledge graphs to improve the accuracy of language models for question answering on structured SQL databases. They explain the creation of a knowledge graph from a data warehouse, evaluate the effectiveness of knowledge graphs in improving question answering accuracy, and discuss how to convince organizations to adopt knowledge graphs for improved data exploration. They also highlight the benefits of knowledge graphs, compare RDF and property graphs, and emphasize the importance of improving knowledge graph accuracy and combining knowledge graphs with vector databases.

Max Mergenthaler and Azul Garza Ramirez from Nixtla talk about TimeGPT, a simplified model for time series analysis. They discuss its simplicity, performance, and potential integration with other tools. They also explore the role of expert judgment and the future impact of TimeGPT on forecasting jobs.

Waleed Kadous, Chief Scientist at Anyscale, discusses best practices for building applications leveraging large language models. Topics include heuristics for working with open source models, differences between Code Lama and GitHub Co-pilot, challenges in deploying open source models, using spending data to save data, fine-tuning models in supervised machine learning, and exploring the potential of multimodal models.

CEO of BlockApps and Co-Chair of the Enterprise Ethereum Alliance discusses web3 technologies, transitioning from mining to proof of stake in Ethereum, impact of Proof of Stake and NFTs in the crypto space, intersection of blockchain, AI, and crypto, creation and governance of base models for ML, current happenings in crypto and web3, advantages of blockchain and collapse of VC market

The podcast discusses the evolution of big data and AI technologies, the rise of open source data in the tech industry, the future of AI and machine learning in a decentralized world, simplifying workload and data movement across cloud and on-prem environments, challenges in data management, and the power of networking in open source data.

Malte Pietsch, co-founder & CTO of Deepset, discusses the importance of orchestration frameworks for LLM applications, the usage patterns of the Haystack framework, and optimizing RAG applications with metadata and knowledge graphs. They also explore the evolution of data engineering pipelines, real-time indexing, and the highlights and features of Haystack 2.0.

The hosts analyze takeaways from the inaugural AI conference in San Francisco, discussing the importance of empirical evidence. Experimenting and iterating in AI leads to improved results. The rise of open source and custom foundation models in AI is explored. The use of ensembles in machine learning and highlights from the AI conference are discussed, including generative AI for speech.

Guest Semih Salihoglu, co-creator of Kuzu, discusses the concept of a property graph, differences between property graphs and RDF in graph databases, the need for switching databases, the design and storage techniques of Kuzu, integration with other programming languages, advantages of DuckDB, and compatibility and streaming in real time.

Philipp Moritz and Goku Mohandas of Anyscale discuss retrieval augmented generation (RAG) systems, challenges in evaluation, labeling and classification strategies, optimizing model inference, online software stack, and hyperparameter search in evaluation runs.

Researchers Bill Marcellino and Nathan Beauchamp-Mustafaga discuss the rise of generative AI and its impact on social media manipulation. They explore the use of generative AI for political and security purposes, motivations of nation-state actors, technology asymmetry, scale and control of information propagation, and combatting manipulative content on social media.