Cover image

One Quarter of the Questions, but Not the Same Retrieval

TL;DR for operators Hierarchical question-based embeddings offer two distinct ways to manage interpretability: use fewer semantic questions in the representation, or summarize detailed questions through broader explanations. Across 20 benchmark datasets, a representation with 3,462 questions retains nearly all the semantic similarity performance of one with 13,914 questions, but its average retrieval score falls more substantially. The optimal level of detail therefore depends on what the embedding must accomplish. ...

November 15, 2026 · 8 min · Zelina
Cover image

The Facts Didn't Make It to the Decision

TL;DR for operators An AI team can possess every fact needed for a decision and still reach the wrong conclusion. TeamGym, a benchmark for distributed-information decision-making, reports an average 22.9-percentage-point loss in decision success when necessary facts are distributed among agents instead of given to one agent. Requiring those agents to retrieve the facts from source documents introduces a further 19.9-point loss.1 ...

November 7, 2026 · 10 min · Zelina
Cover image

A Better ICD Shortlist Is Not Yet a Better Coding Workflow

TL;DR for operators Hospital coding teams must identify appropriate diagnosis codes from clinical notes, laboratory results, and admission records. The first automation decision is whether software should predict codes directly or retrieve a shortlist that a coding professional can examine. Bohra and Bohra’s Multimodal Dual-Encoder Retrieval for Automated ICD Coding1 provides evidence for the retrieval approach. On MIMIC-III, its model achieves 66.75% Precision@3 versus 62.12% for the strongest multimodal classification baseline, and 62.44% versus 59.73% Micro-F1@8. A separate LLM stage improves the ordering of retrieved candidates, raising NDCG@8 from 65% to 71% on 2,000 human-annotated cases. ...

November 4, 2026 · 7 min · Zelina
Cover image

Before Search Comes Normalization

TL;DR for operators A voice product cannot safely assume that Arabic speech recognition output and canonical Quranic text are already comparable strings. The paper constructs a 2,290-pair mapping between Uthmani word forms and Standard Arabic, then embeds that resource in a deterministic normalization and retrieval system.1 The practical result is substantial but bounded. After normalization, 5,669 of 6,236 verses—90.9%—become identical across the two written forms. The validator passes 122 of 124 cases in its released evaluation suite. Yet both failures come from the same architectural choice: an ASR substitution can accidentally create an exact match for the wrong verse, and the system accepts that candidate before comparing alternatives. ...

October 16, 2026 · 7 min · Zelina
Cover image

Search With Vision, Answer With Text? What IRPAPERS Changes About Scientific RAG

TL;DR for operators A team building RAG over scientific PDFs has at least two separate decisions to make: how a page should be found, and what representation of that page the answering model should read. Those decisions do not receive the same answer. In the IRPAPERS benchmark,1 text and image retrieval recover partially different correct pages. Fusing their signals raises open-source retrieval to 49% Recall@1 and 95% Recall@20, above either modality alone. But once retrieved material is passed to GPT-4.1 for question answering, OCR text is substantially stronger: TextRAG scores 0.82 alignment at five retrieved pages versus 0.71 for ImageRAG. ...

September 14, 2026 · 7 min · Zelina
Cover image

K-Means, K-Gone: Sparse Coding and the Retrieval Bottleneck

Indexing is where many retrieval systems quietly become expensive. The demo looks harmless: upload documents, create embeddings, ask questions, receive answers with citations. Then the corpus starts behaving like a real business corpus. Policies change. Product pages are rewritten. Compliance documents are replaced. Support tickets arrive every hour. The retrieval layer must keep up, and suddenly the glamorous RAG stack is waiting for the plumbing to rebuild itself. As usual, the least photogenic component is the one holding the invoice. ...

June 2, 2026 · 21 min · Zelina
Cover image

RAG and the Art of Not Dropping the Answer

RAG and the Art of Not Dropping the Answer A RAG team usually starts with a familiar ambition: make the retrieved context smarter. The raw document feels too long. The search snippet feels too primitive. The page structure looks messy. A query-focused summary sounds more elegant. A proposition list sounds more machine-readable. A paraphrase from a strong LLM sounds, at least cosmetically, like an upgrade. So the team builds another representation layer between retrieval and generation, hoping the model will reward the extra sophistication. ...

June 2, 2026 · 16 min · Zelina
Cover image

Read the Receipt: Why RAG Should Highlight Before It Answers

Search looks easy until someone asks where the answer actually came from. A researcher types a rough query into a literature assistant. The system retrieves several papers, writes a fluent answer, and appends citations. Everyone relaxes a little. The citation tag has done its small administrative magic. The answer now looks grounded. ...

May 30, 2026 · 15 min · Zelina
Cover image

Wide Thinking, Narrow Context: Why InfoSeeker Rewrites the Economics of AI Search

A spreadsheet is a cruel test of artificial intelligence. Not the toy spreadsheet used in demos, with six rows, three columns, and a suspiciously cooperative universe. I mean the kind of table a real analyst asks for: every qualifying supplier in a region, every product SKU released over a decade, every regulatory filing matching a narrow condition, every competitor with exact addresses, dates, sources, and no missing cells because apparently human suffering needs columns. ...

April 6, 2026 · 16 min · Zelina
Cover image

No More ‘Trust Me, Bro’: Statistical Parsing Meets Verifiable Reasoning

AI systems are very good at saying things. This is both the miracle and the invoice. In enterprise settings, the sentence itself is rarely the final product. A compliance officer does not only want an answer about whether a clause violates policy. A credit analyst does not only want a summary of why a borrower looks risky. A procurement team does not only want a generated explanation of why Vendor A seems eligible. They want to know what the system used, which rule it applied, where the uncertainty sits, and whether the conclusion survives when the evidence changes. ...

February 13, 2026 · 17 min · Zelina