AMiner

Tsinghua-born academic search & knowledge-graph platform — one of the largest scholarly entity graphs, operated by Zhipu and now blending LLMs with structured academic knowledge.

1. Core Product / Service

AMiner (aminer.cn) is an academic search, mining, and expert-profiling platform that originated in Tsinghua University's Knowledge Engineering Group (KEG). It indexes scholarly literature, researchers, institutions, and citation relationships into a large knowledge graph — nodes (papers / authors / venues / institutions), edges (citations / co-authorship / affiliation), and attributes (impact metrics, research topics) [1].

Core capabilities:

• Academic search across a large paper/researcher corpus, with expert profiles, co-author networks, and citation analytics.
• Knowledge-graph infrastructure — a structured scholarly graph used for expert discovery, tech scouting, and institution benchmarking.
• Graph + LLM synergy — the current product direction: structured graph knowledge reduces LLM hallucination and improves explainability, while LLMs auto-complete missing graph relations. Graph RAG (retrieval-augmented generation over the graph) is positioned as the bridge [1].

AMiner is operated by Beijing Zhipu Huazhang Technology Co. (智谱华章), the same corporate vehicle behind zhipu|Zhipu AI, sharing the Tsinghua KEG lineage.

2. Target Users & Pain Points

• Researchers & universities — literature discovery, collaborator finding, expert profiling.
• Enterprises & investors — tech scouting, talent and institution due-diligence via the scholar graph.
• LLM builders — structured academic knowledge as a grounding source to cut hallucination and add verifiable citations.

Pain solved: raw literature search returns documents; AMiner's value is the graph — who works on what, who cites whom, how institutions cluster — which is hard to reconstruct from keyword search alone.

3. Competitive Landscape

AMiner's differentiation is the expert/social-graph layer (co-author networks, institutional clustering) built over years of Tsinghua KEG research, plus its tight coupling to Zhipu's GLM stack for the graph+LLM synergy play [1].

4. Unique Observations

• Same lineage as Zhipu. AMiner and zhipu|Zhipu AI both trace to Tsinghua's KEG lab (anchored by Prof. Tang Jie / 唐杰), which makes AMiner one of the few academic-search products with a direct frontier-LLM sibling. The commercial link (both under 智谱华章) means AMiner can feed structured academic knowledge into GLM-family models and vice versa.
• Knowledge-graph as grounding, not just retrieval. The Graph RAG framing is the interesting bet: vector retrieval is fast but factually loose; the structured graph trades some token cost (local research reports 20–30% more than pure vector) for lower hallucination (15%) and better explainability [1]. This is a concrete, defensible value prop in a market where LLM hallucination is the top blocker for academic/enterprise adoption.
• Research program, not just a feature. AMiner Research Labs' "academic graph + LLM" white paper (《学术图谱+大模型协同》) and signed joint research with ~5 university labs signal the graph+model agenda is being pushed as a research program [1].

5. Financials / Funding

No standalone funding disclosed — AMiner is a lab-originated platform operated within zhipu|Zhipu (智谱华章), which listed on the Hong Kong Stock Exchange in January 2026. AMiner's economics are folded into Zhipu's academic/enterprise business rather than tracked as an independent venture.

6. People & Relationships

• Academic origin: Tsinghua University Knowledge Engineering Group (KEG); Prof. Tang Jie (唐杰) is the academic anchor and AMiner's research lead.
• Operator: zhipu|Zhipu AI (北京智谱华章科技) — the same corporate entity behind the GLM model family and BigModel API.
• Competitors: Google Scholar, Semantic Scholar, DBLP, OpenAlex (plain text — no dedicated pages in this wiki).