Reading
The research team integrated the hierarchical semantics of the ECLASS standard into embedding retrieval, combining dense retrieval and re-ranking technologies to achieve a 94.3% Hit@5 rate in industrial electronic component search tasks, far exceeding the traditional BM25 method.
Section 01
The research team integrated the hierarchical semantics of the ECLASS standard into embedding retrieval, combining dense retrieval and re-ranking technologies to achieve a 94.3% Hit@5 rate in industrial electronic component search tasks—far exceeding the traditional BM25 method (31.4%)—and providing a more intelligent solution for industrial search scenarios.
Section 02
In the era of industrial automation and intelligent manufacturing, engineers need to quickly find suitable components from massive product catalogs, but traditional search faces vocabulary mismatch issues: there is a gap between human natural language queries and structured product attribute descriptions (e.g., manufacturer, model, parameters), and lexical retrieval methods like BM25 have poor performance.
The semantic understanding capabilities of large language models are expected to bridge this gap. Dense retrieval captures relevance by mapping text to a semantic vector space, but the short length of product descriptions and dense professional terminology place higher demands on embedding quality.
Section 03
ECLASS is an international industrial product classification standard that provides a hierarchical semantic structure (e.g., "Electronic Components > Semiconductor Devices > Microcontrollers") and standardized attribute definitions. Its rich semantics have not been fully utilized by traditional systems.
Section 04
Section 05
Section 06
Section 07
ECLASS-enhanced semantic retrieval demonstrates the power of combining domain knowledge with general AI, achieving a leap in Hit@5 from 31.4% to 94.3%. It suggests that breakthroughs in AI applications require in-depth understanding of domain characteristics and the conversion of implicit knowledge into machine-computable forms.
Keep going with more reads from the same topic.
Splinter is a minimalist, high-performance key-value (KV) and vector storage system enabling zero-latency inter-process communication via shared memory and atomic operations. With only 766 lines of core code, it supports millions of operations per second and 768-dimensional vector storage, offering a new architectural approach for local LLM inference and data-intensive applications.
Folkering OS is the world's first AI-native bare-metal operating system, entirely written in Rust no_std without relying on Linux, POSIX, or libc. It can generate commands from scratch, compile them into WASM, and run them in 10 seconds, achieving true self-evolution.
This project explores how to use Large Language Models (LLMs) to detect logical vulnerabilities in Ethereum smart contracts, especially business logic flaws that are difficult for traditional static analysis tools to capture, such as economic manipulation and execution order errors.
Cross-generational hardware testing based on NVIDIA A10G and H100 GPUs, comparing and analyzing the throughput, latency, and concurrent scalability of three mainstream LLM inference frameworks. SGLang achieves a 3.4x performance advantage over vLLM on H100, while Ollama faces architectural bottlenecks in high-concurrency scenarios.