Various speakers: SISAP 2026 Indexing Challenge accepted papers

The SISAP 2026 Indexing Challenge (https://sisap-challenges.github.io/2026/index.html) —part of the SISAP 2026 (Similarity Search and Applications) conference— focuses on advancing state-of-the-art similarity search and indexing techniques across diverse data
modalities and scalability benchmarks. This session highlights innovative approaches to address the complex computational and structural bottlenecks inherent in high-dimensional and large-scale retrieval tasks. To showcase these developments, four short technical presentations will be delivered, each introducing original and efficient solutions tailored to the challenge's distinct tasks. These are:

* Luis Solsona Sánchez and Luis Fernando Chávez Quevedo: Memory-Bounded Approximate Nearest Neighbour Search for Dense, MIPS, and Sparse Vectors: SISAP 2026 Indexing Challenge

* Aftab Ahmed Choudhry: Hardware-Sympathetic Optimization for High-Dimensional Vector Search

* Eneko Sabaté Iturgaiz: NMF-IVF: Learning Latent Semantic Topics for Approximate Retrieval

* Sergi Guimerà Roig and Iván Sánchez Gala: PiPNN for k-NN Graph Construction: Inverse-Size Merging and Diversified Beam Search

The talks will cover key methodological breakthroughs, experimental evaluations, and practical insights gained from tackling real-world indexing constraints.

This seminar will conclude with a short talk by Àlex Font, former MIRI student, where he will present the work that was accepted as a full regular paper at the SISAP 2026 main conference (https://sisap.org/2026).

* Àlex Font, Conrado Martínez, Mahmudur R. Hera and David Koslicki: A Detailed Analysis of MaxGeomHash Sketches