Yanez
Multimodal
Inorganic Identities
Datasets
The main goal of Yanez MIIDs is to generate synthetic identities that can be used to test and train anti-fraud models and identity recognition algorithms. An inorganic identity can be an identity that is entirely made up with fake attributes, or it can be one that incorporates aspects of real identities. Different flavors of identities are required for different use cases.
Roadmap
Phase 1: Threat Scenario Query Execution & Initial Deployment
May 1, 2025
-
Deploy the Yanez subnet on Bittensor’s mainnet after thorough local testing and testnet validation.
-
Introduce the Threat Scenario Query System, enabling validators to reward miners based on known threat scenarios.
-
Develop execution vectors targeting name-based evasion tactics, including:
-
Phonetic manipulations
-
Orthographic variations
-
Rule-based alterations
-
-
Deploy initial dataset comprising synthetic name variations for threat modeling.
Phase 2: Threat Scenario Query Expansion & Miner Contributions
June 12, 2025
-
Expand Threat Scenario Query System to allow proposed (unknown) threat scenarios from miners.
-
Introduce mechanisms for miners to submit new evasion tactics, including:
-
-
Nicknames-based threats
-
Transliteration-based alterations
-
Middle name manipulations
-
-
Improve the validator scoring system, ensuring:
-
Execution vector diversity in submitted queries.
-
Penalty enforcement for repetitive or low-value scenarios.
-
Phase 3: Threat Scenario Expansion – Location-Based Identity Manipulation
August 1, 2025
-
Expand threat scenarios to include location-based obfuscation techniques.
-
Establish known location-based execution vectors, including:
-
IP address manipulation
-
Geographic inconsistencies
-
Misspelled or altered sanctioned regions
-
-
Develop organic queries for Yanez Compliance, analyzing:
-
High-risk jurisdiction evasion patterns.
-
Transaction laundering through location obfuscation.
-
-
Introduce miner-driven contributions to propose and validate unknown location-based attack vectors.
Phase 4: AML Ecosystem Integration & Threat Scenarios Evaluation
September 15, 2025
-
Fully integrate the Yanez subnet into financial crime testing workflows, making its subnet accessible for external validation and industry collaboration.
-
Enable external stakeholders (financial institutions, compliance teams, regulators) to interact with the Yanez subnet by:
- Using the generated synthetic data as-is for direct evaluation.
-
Making organic compliance queries to test against real-world threat scenarios dynamically.
-
Expand the database of known execution vectors, allowing continuous refinement and contributions from external entities.
1
2
4
🔹 Technical Value:
-
Move to mainnet: Enables wider audience of miners and validators to join the subnet and build the foundation for the subnet’s ultimate goals.
-
Known Threat Scenarios: Establishes a foundation for understanding evasion tactics and generating execution data that accurately tests sanctions compliance screening.
-
Execution Vectors Normalization: Leaning threat scenarios distribution patterns.
💼 Business Value:
-
Regulatory Compliance & Audit Readiness: Demonstrating solid detection capabilities for name variations ensures meeting auditors expectations.
-
Improved Screening Accuracy: Helps compliance teams in achieving operational efficiencies and effectiveness by focusing on variations that matter and reducing false positives.
-
Market Differentiation: Positions Yanez Compliance as a cutting-edge AI-driven testing and automation platform, accelerating adoption in the financial sector.
🔹 Technical Value:
-
Proposed Threat Scenarios: Enables community-driven innovation, allowing miners to enhance the threat library.
-
Post-Evaluation System: Establishes a structured validation process to assess new threat scenarios.
-
Learning from Miner Contributions: Improves the coverage of name-based evasion tactics that traditional screening models may not be capable of detecting.
💼 Business Value:
-
Adaptive Threat Intelligence: Enables financial institutions to stay ahead of evolving evasion tactics, reducing potential compliance gaps and fines, and more importantly gaps that can be exploited by nefarious actors.
-
Competitive Advantage: Yanez Compliance becomes a crowdsourced intelligence platform, increasing the chances for financial institutions to move ahead in the threat race.
-
Revenue Potential: Institutions could license the validated threat scenario datasets for training and thus enhancing their detection models.
🔹 Technical Value:
-
Detect Location-Based Threats: Improves location-based risk screening for sanctions compliance.
-
Leverage Miner Insights: Harnesses crowdsourced knowledge to identify regional evasion tactics.
💼 Business Value:
-
Global Risk Mitigation: Helps detect sanctions evasion attempts based on location.
-
New Market Penetration: Expands Yanez AI’s relevance to global AML community.
🔹 Technical Value:
-
Open-Access Threat Scenario Validation: Organizations can directly interact with Yanez AI to test their screening models.
-
AML Ecosystem: Establishes an incentive-driven compliance ecosystem, encouraging high-quality data contributions.
-
Bid-Based Threat Scenario Challenges: External compliance teams can submit validation challenges of new attack vectors tailored to emerging financial crime risks.
💼 Business Value:
-
Global Industry Collaboration: Yanez Compliance opens its subnet for AML vendors, compliance officers, and regulators to validate risk scenarios in real-time.
-
Data Monetization & Licensing: Enables financial institutions and fintech companies to leverage Yanez’s risk datasets for model training and compliance validation.
-
Proof of Compliance Innovation: Establishes Yanez as an industry leader by offering an AI-driven, miner-validated compliance model that evolves dynamically with real-world financial crime trends.
Phase 5: Biometric Data Integration Q1 2026
-
Incorporate facial biometric data to enhance identity realism.
-
Introduce mechanisms for detecting fraudulent biometric identities.
5
Phase 6: Synthetic Document Generation Q2 2026
-
Develop capabilities to generate synthetic documents associated with identities.
-
Validate document authenticity to counteract fraud attempts.
6
Phase 7: Synthetic Digital Presence Creation Q3 2026
-
Simulate online behaviors and interactions linked to synthetic identities.
-
Model evasion tactics for digital footprint analysis.
7
Phase 8: Financial Transaction Simulation Q4 2026
-
Enable digital personas to perform financial transactions within virtual environments.
-
Assess compliance system response to synthetic financial activities.
8
Phase 9: 3D Model Avatars for Identity Representation Q2 2027
-
Create 3D avatars to visually represent synthetic identities.
-
Develop variations reflecting different demographic and geographic attributes.
9
Phase 10: Voice Feature Integration --
-
Introduce voice synthesis capabilities to add realism to synthetic identities.
-
Develop safeguards against voice-based fraud.
10
Phase 11: Conversational AI Integration ---
-
Enable avatars to interact dynamically using conversational AI.
-
Train models for multilingual communication in screening systems.
11
Final Phase: Unified Identity Representation & ML Training ---
-
Train a robust unified model for identity recognition and screening.
-
Leverage AI-driven proof-of-life models to verify real vs. synthetic identities.
-
Establish a decentralized platform for organizations to contribute and validate identity screening topics.
12
