Relevance in the Age of Generative Search - Haystack US 2023 Keynote
Explore strategies for integrating generative AI and language models into search applications, balancing accuracy and relevance while leveraging new capabilities in the evolving search landscape.
Creating Representative Query Sets for Offline Evaluation
Learn to construct representative query sets for offline evaluation in AI/ML experimentation, enhancing understanding of algorithm impacts and streamlining A/B testing processes.
Learning to Hybrid Search: Combining BM25, Neural Embeddings, and Customer Behavior
Combine BM25, neural embeddings, and customer behavior for ultimate ranking. Compare traditional, LTR, and neural search approaches in e-commerce using Amazon's ESCI dataset.
Dive into NLP with the Elastic Stack
Explore practical NLP applications using Elasticsearch, including semantic vectorization, semantic search, named entity recognition, multimodal search, and automatic classification.
Vector Search for Clinical Decisions
Explore EBSCO's journey in implementing vector search for clinical decision support, addressing challenges in precision and recall for complex medical queries.
Better Semantic Search with Hybrid Sparse-Dense Search - Haystack US 2023
Explore hybrid search techniques combining sparse and dense embedding models to enhance semantic search accuracy and effectiveness in various scenarios.
Enterprise Search Relevance at Box: Simplicity
Explore Box's approach to delivering relevant search results across diverse enterprises using Solr and TensorFlow Ranking, balancing simplicity with privacy and security.
Building Recommendation Systems with Vector Search - Women of Search
Explore building recommendation systems using vector search, focusing on Ref2Vec for user representation and bipartite graphs for user-product connections. Learn to enhance e-commerce experiences through advanced search techniques.
Bias Identification and Mitigation in AI Systems
Explore techniques for identifying and mitigating bias in AI systems with NLP expert Arnault Gombert.
AI Search and Generative Answers with Open-Source NucliaDB
Explore AI-powered search engines with Nuclia's API, featuring generative answers and open-source NucliaDB technology.
Reciprocal Rank Fusion (RRF) - How to Stop Worrying about Boosting - Lecture
Explore Reciprocal Rank Fusion (RRF) for combining search results, its advantages over boosting, and implementation in various systems. Learn from Philipp Krenn's expertise in search technology.
Shedding Light on Positional Bias: Strategies for Mitigation in Search Ranking - Haystack EU 2023
Explore strategies to mitigate positional bias in search rankings, improving relevance and fairness in ML-based systems for better user experience and search quality.
Beyond the Known: Exploratory and Diversity Search with Vector Embeddings - Haystack EU 2023
Explore innovative strategies for exploratory and diverse search using vector embeddings, enhancing user experiences beyond traditional search paradigms.
Navigating Neural Search: Avoiding Common Pitfalls
Explore neural search, text embedding models, and vector retrieval. Learn to choose models based on language, task alignment, accuracy, and cost-effectiveness. Understand potential pitfalls in embedding retrieval.
Mastering Hybrid Search: Blending Classic Ranking with Vector Search - Haystack EU 2023
Explore hybrid search systems combining classic ranking and vector search. Learn best practices for enhancing relevancy, managing performance-accuracy tradeoffs, and scaling to production.
The Art of Hypothesis-Driven Development of ML-Powered Search - Haystack EU 2023
Explore ML-powered search development through hypothesis-driven approaches. Learn from Uzum Market's journey in building a large-scale search system for e-commerce, focusing on metrics, evaluation, and AB testing.
Using Vector Databases to Scale Multimodal Embeddings, Retrieval and Generation
Explore scaling multimodal embeddings, retrieval, and generation using vector databases. Learn cross-modal search techniques and guide large language models with real-time retrieval.
Combining Inverted and ANN Indexes for Efficient Hybrid Search at Scale
Explore challenges and solutions for integrating inverted and ANN indexes in search engines, optimizing hybrid search systems for efficiency and scalability.
Evaluating Embedding-Based Retrieval Beyond Historical Search Results
Explore methods to evaluate embedding-based retrieval quality using organizational knowledge, improving relevance and expanding recall for e-commerce search beyond historical data.
Can Search Results be Fair? - Haystack On Tour Austin
Explore fairness in search results with NIST Scientist Emeritus Ellen Voorhees, discussing challenges and potential solutions in information retrieval systems.
All Vector Search is Hybrid Search
Explore the benefits and challenges of hybrid search systems, combining vector search with traditional methods for improved accuracy and real-time updates in search applications.