Verified Record & Original Data

Research Record, Patents, and Survey Data

Last updated: All counts computed from public filings and committed datasets on amirshachar.com

Summary

Amir Shachar is a fraud scientist, patented inventor, and founder and Chief Scientist of Riskernel. He previously served as Chief AI Scientist at Skyhawk Security and Head of Fraud Data Science at NICE Actimize, where he received the 2019 CEO Excellence Award for patented commercial deployment of federated learning across Tier-1 financial institutions. He is the named inventor on 14 patents (9 granted in the US, 5 pending) and author of the Algogens framework and Semi-discrete Calculus.

1. Verified Public Record Summary

This table consolidates verifiable public counts across Amir Shachar's career, patent filings, arXiv preprints, literature survey datasets, press coverage, and university teaching.

Category Verified Figure Scope & Breakdown Primary Source
Named inventor patents (total) 14 (9 granted US, 5 pending) 2014 to 2026; NICE Ltd. (9), Skyhawk Security Ltd. (3), Outbrain Inc. (1), Riskernel provisional (1) /patents/
Public arXiv preprints 3 preprints (incl. 560-page Algogens monograph) arXiv:2403.01426 (2024, 560 pp.), arXiv:1012.5751 (2010), arXiv:1005.1418 (2010) /publications/
Books and manuscripts under review 1 self-published book, 2 manuscripts under review Introduction to Semi-Discrete Calculus (2022); An Order Sensor (2026); Three Ways to Be Increasing at a Point (2026) /detachment-library/
Detachment Library catalogued entries 808 total (802 survey + 6 author works) 4 branches, 41 fields, 166 panels, 1,050 references (1994 to 2021) /detachment-library/
Curated hybrid AI systems index 100 hybrid AI systems Curated Algogens hybrid AI index referenced from /algogens/ and riskernel.com/about/ /algogens/
On-site research papers and technical notes 8 pages Algogens, Semi-discrete Calculus, Trendland, TrenDescent, Inquiry Policy Learning, Semi-supervised Reasoning, Natural Sciences, AI & Security /publications/
Fraud data science essays (canonical on riskernel.com) 4 posts (published 2026-11-12) Alert budgets, propose-then-verify stacks, fraud rings and mule networks, delayed and censored labels /blog/
Press coverage and podcast interviews 6 press articles, 2 Hebrew podcasts SC Media, TechTarget, startuphub.ai, CSO, Dark Reading, Yahoo Finance (2023 to 2024); Little Big Science and ExplAInable (2024) /media/
University teaching record (Technion) 4 courses (2008 to 2010) TA for Calculus 1, Advanced Calculus, Linear Algebra 1, and Image Processing /teaching/
Documented industry recognition 2019 NICE Actimize CEO Excellence Award Awarded for patented commercial deployment of federated learning across Tier-1 financial institutions riskernel.com/about/

Last updated: . Methodology and provenance: Patent counts come from the 14 public records indexed on /patents/ (verified against Google Patents and Riskernel). Survey counts are computed at build time from src/data/detachment-library.json, src/data/detachment-research.json, and src/data/trendland.json. Career and award statements come from public records on riskernel.com/about/. Private client names, engagement fees, and NDA metrics are excluded by policy.

2. Patent Portfolio Breakdown (14 Patents)

Amir Shachar is a named inventor on 14 patents spanning 2014 to 2026: 9 granted in the United States and 5 pending applications.

Organization / Assignee Granted (US) Pending Total Years Technical Focus
NICE Ltd. 5 4 9 2020 to 2023 Federated learning, extreme class imbalance, quantile summaries, model testing, decision rules, semi-discrete meta-learning
Skyhawk Security Ltd. 3 0 3 2023 Cloud security posture management, self-supervised threat detection, anticipated attacker pathway evaluation
Outbrain Inc. 1 0 1 2014 Content title user engagement optimization
Riskernel 0 1 1 2026 Algogens fraud detection logic proposal, deterministic validation, and federated sharing (US provisional)
Total 9 5 14 2014 to 2026 Machine learning, privacy-preserving AI, federated learning, cloud security, fraud detection

Complete 14-Patent Reference Ledger

Identifier Status Assignee / Context Year Title
US provisional 64/197,364 Pending Riskernel 2026 How the Riskernel Engine Proposes, Checks, Shares and Runs Fraud Detection Logic
US 12,634,298 B2 Granted (US) Skyhawk Security Ltd. 2023 Security Posture Management Methods and Systems Using Threat Detection and Response Data
US 12,626,151 B2 Granted (US) NICE Ltd. 2023 Method and System for Testing a Classification Machine Learning (ML) Model of a Tenant of a Service Provider, in a Cloud-Based Environment
US 12,541,721 B2 Granted (US) NICE Ltd. 2022 Method for Extreme Class Imbalance Within Fraud Detection
US 12,481,890 B2 Granted (US) NICE Ltd. 2020 Systems and Methods for Applying Semi-discrete Calculus to Meta Machine Learning
US 12,476,990 B2 Granted (US) Skyhawk Security Ltd. 2023 Method and System for Early Detection of Malicious Behavior Based Using Self-Supervised Learning
US 12,425,439 B2 Granted (US) Skyhawk Security Ltd. 2023 Methods and Systems for Threat Evaluation Based on Anticipated Pathways of Malicious Actors
US 12,124,933 B2 Granted (US) NICE Ltd. 2020 Artificial Intelligence System for Anomaly Detection in Transaction Data Sets
US 12,118,558 B2 Granted (US) NICE Ltd. 2021 Estimating Quantile Values for Reduced Memory and/or Storage Utilization and Faster Processing Time in Fraud Detection Systems
US 10,607,253 B1 Granted (US) Outbrain Inc. 2014 Content Title User Engagement Optimization
US 2024/0394713 A1 Pending NICE Ltd. 2023 Systems and Methods for Programmatically Generating Decision Rules for Alert Detection Engines Using Machine Learning Operations
US 2023/0267468 A1 Pending NICE Ltd. 2022 Unsupervised Clustered Explanation-based Feature Selection Using Transfer Learning for Low Fraud Scenario
US 2023/0237494 A1 Pending NICE Ltd. 2022 Systems and Methods for Automatically Creating Machine Learned Fraud Detection Models
US 2023/0306429 Pending NICE Ltd. 2022 Method for Maintaining Ethical Artificial Intelligence (AI)

Last updated: . Methodology and provenance: Generated directly from the structured patent entries in src/content/pages/patents.md, with external links pointing to Google Patents, Justia Patents, and Riskernel's public provisional filing summary (US provisional 64/197,364).

3. Detachment Library and Trendland Survey Data

The Detachment Library indexes 808 total entries: 802 published applications from the Trendland literature survey across 4 scientific branches and 41 fields, plus 6 author works by Amir Shachar.

Scientific Branch Survey Entries Share of Survey Sub-fields Largest Top-Level Fields
Science 394 49.1% 19 Natural Sciences, Social sciences, Data Science
Engineering 300 37.4% 11 Computer science, Electrical engineering, Mechanical Engineering
Mathematics 78 9.7% 11 Statistics, Mathematical Optimization, Mathematical Analysis
Science Education 30 3.7% 0 Direct branch entries (no sub-fields)
Total Literature Survey 802 100.0% 41 166 tree panels, 1,050 resolved references, 1,305 math spans (1994 to 2021; 797 dated cards)

Last updated: . Methodology and provenance: Every count is computed automatically during site build from src/data/detachment-library.json and src/data/trendland.json, with zero unresolved reference gaps (1,050 of 1,050 references verified). Six author works from src/data/detachment-research.json bring the total library size to 808 entries.

4. Research Papers, arXiv Preprints, and Fraud Writing

Complete index of Amir Shachar's public arXiv preprints, self-published book, manuscripts under review, on-site research essays, and canonical fraud data science essays hosted on Riskernel.

Work Type Date / Status Length / Identifier
Introduction to Algogens: Integrating Generative AI with Algorithmic Frameworks arXiv monograph 2024-03-03 arXiv:2403.01426 · 560 pages
Introduction to Semi-discrete Calculus arXiv preprint 2010-12-28 arXiv:1012.5751
On a Relation Between the Integral Image Algorithm and Calculus arXiv preprint 2010 arXiv:1005.1418
Introduction to Semi-Discrete Calculus (Book) Self-published book 2022 Public entry point on /semi-discrete-calculus/
An Order Sensor: Diagnosing Folds, Kinks, and Multiplicity in Differentiable Programs Manuscript Under review (2026) Venue unnamed during review
Three Ways to Be Increasing at a Point Manuscript Under review (2026) Venue unnamed during review
Algogens: Generative AI Exploring Inside Algorithmic Guide Rails On-site research page Updated 2026-11-20 4,556 body words
A Trend Operator: Theory and AI Applications On-site research page Updated 2026-11-13 14,435 body words
Exploring the Trend of Exploiting the Trend On-site research page Updated 2026-11-12 388 body words
Semi-Supervised Reasoning: Insights from Unlabeled Data On-site research page Updated 2026-11-12 1,402 body words
TrenDescent On-site research page Updated 2026-11-12 776 body words
Inquiry Policy Learning: Evaluator Repair and Epistemic Control for Self-Evolving AI Systems On-site research page Updated 2026-05-06 7,803 body words
Leveraging AI Security to Counter AI Threats On-site research page Updated 2024-02-01 PDF paper (9.7 MB)
Calculating Trends in Natural Sciences On-site research page Updated 2023-09-18 2,562 body words
Alert budgets, not AUC: judging a fraud model in the units that matter Fraud data science essay 2026-11-12 Canonical on riskernel.com
Propose, then verify: where a language model belongs in a fraud stack Fraud data science essay 2026-11-12 Canonical on riskernel.com
Fraud rings and mule networks: what a graph adds, and what it cannot decide Fraud data science essay 2026-11-12 Canonical on riskernel.com
The label you get is not the label you need Fraud data science essay 2026-11-12 Canonical on riskernel.com

Last updated: . Methodology and provenance: Word counts for on-site essays are calculated directly from Markdown/MDX source files excluding code fences and raw TeX spans. arXiv dates and page count (560 pages for arXiv:2403.01426) come from arXiv metadata archived in design/rebrand/SOURCE-PACK.md. The four fraud data science essays were moved canonically to riskernel.com on 2026-11-19 with 301 redirects from amirshachar.com.

5. Frequently Asked Questions

Where do the counts on the Research Record and Data page come from?
Every count on this page is computed directly from committed repository data and public filings: the 14 patent entries linked to Google Patents and Riskernel on /patents/, the 802 literature survey entries and 1,050 references in Trendland and the Detachment Library, the 3 public arXiv preprints and 2 manuscripts under review, the 4 fraud writing essays on riskernel.com, and the press, interview, and Technion teaching archives.
How are Amir Shachar's 14 patents distributed across organizations?
Nine patents are assigned to NICE Ltd. (5 granted US patents and 4 pending applications, 2020 to 2023), three granted US patents are assigned to Skyhawk Security Ltd. (2023), one granted US patent is assigned to Outbrain Inc. (2014), and one sole-inventor US provisional application (64/197,364, 2026) covers the Riskernel engine.
What does the 808-entry Detachment Library measure?
The Detachment Library catalogues 802 published applications of one-sided local trend signs from the Trendland literature survey across 4 branches (Science: 394, Engineering: 300, Mathematics: 78, Science Education: 30), 41 fields, 166 panels, and 1,050 references spanning 1994 to 2021, alongside 6 author works by Amir Shachar (2 arXiv preprints, 1 granted patent, 1 self-published book, and 2 manuscripts under review).
Why are client names, engagement pricing, and private model benchmarks omitted?
Engagements run under NDA, so client names, commercial fees, and private production metrics are intentionally excluded from amirshachar.com. Every number published on this site can be traced to a public patent record, an arXiv preprint, or an inspectable dataset on this domain.