Fno44 commited on
Commit
ec49d99
·
verified ·
1 Parent(s): 5478979

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +58 -0
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ - ja
5
+ tags:
6
+ - graph-ml
7
+ - technology-assessment
8
+ - due-diligence
9
+ - world-models
10
+ - ecosystem-analysis
11
+ library_name: transformers
12
+ pipeline_tag: graph-ml
13
+ ---
14
+
15
+ # Ecosystem Analysis AI — WARP DD Model 01
16
+
17
+ ## Overview
18
+
19
+ Ecosystem Analysis AI is the first of three proprietary AI models powering [WARP DD](https://www.linedotjp.com), an AI-powered technology due diligence platform by [LINEdot., Inc.](https://huggingface.co/linedot)
20
+
21
+ This model captures the relational structure of technology ecosystems — mapping how technologies, research, and markets interconnect and evolve.
22
+
23
+ ## Architecture
24
+
25
+ - **Foundation:** Graph Neural Network (GNN)
26
+ - **Approach:** Models the tripartite relationships between technology, research, and market as a dynamic graph
27
+ - **Input:** Global technology data sources (papers, patents, GitHub, funding data, etc.)
28
+ - **Output:** Ecosystem structure analysis, relationship mapping, influence scoring
29
+
30
+ ## What It Does
31
+
32
+ - Maps the entire technology ecosystem as a dynamic graph
33
+ - Identifies hidden connections between technologies, companies, and research domains
34
+ - Detects emerging clusters and structural shifts in real-time
35
+ - Provides the foundational "world model" that feeds into Tech Profiling AI and De Facto Trajectory AI
36
+
37
+ ## Part of the WARP DD 3-Layer AI Engine
38
+
39
+ | Model | Role |
40
+ |-------|------|
41
+ | **Ecosystem Analysis AI** (this model) | Graph-based relational structure modeling |
42
+ | **Tech Profiling AI** | 6-axis quantitative competitiveness evaluation |
43
+ | **De Facto Trajectory AI** | Future trajectory prediction with causal inference |
44
+
45
+ ## Validation
46
+
47
+ - **89%** overall backtest accuracy across 100 AI/ML companies (2-year validation)
48
+ - **~200 companies** across 7 technology domains validated with real-world data
49
+
50
+ ## Note
51
+
52
+ Model weights and training code are proprietary and not publicly available. This Model Card is provided to document the architecture and capabilities of the model.
53
+
54
+ ## Links
55
+
56
+ - 🌐 [Website](https://www.linedotjp.com)
57
+ - 🐙 [GitHub](https://github.com/linedot-ai)
58
+ - 💼 [LinkedIn](https://www.linkedin.com/company/125043999/)