μ μ κ³Ό μ§λ¨κ³Ό λμ§νΈ μ μ 건κ°μ ν¨λ¬λ€μμ λ°κΎΈλ 4건μ μκ³ λ¦¬μ¦ νΉν. 8κ° λ°μ΄ν° μμ€λ₯Ό λ² μ΄μ§μ μΆλ‘ μΌλ‘ ν΅ν©νκ³ , 18κ° DSM-5-TR μ§λ¨μ λμμ μ°μΆνλ©°, μΈκ°-AI λνμ μ λ’°λλ₯Ό κ°κ΄μ μΌλ‘ μΈ‘μ νλ μμ€ν . Four algorithm patents reshaping the paradigms of psychiatric diagnosis and digital mental health: a system that fuses eight data sources through Bayesian inference, simultaneously yields eighteen DSM-5-TR diagnoses, and objectively measures the reliability of human-AI conversation.
μλ 4건μ νΉνλ λͺ¨λ μ€μ μλνλ μμ€ν (agedlearning.com)μ ꡬνλμ΄ μμΌλ©°, Boston Neuromind LLCμ μμ λ°μ΄ν°λ‘ κ²μ¦λκ³ μμ΅λλ€. ν΅μ¬μ DMDA(Patent 4)μ΄λ©°, λλ¨Έμ§ 3건μ DMDAλ₯Ό λλ¬μΌ 보쑰 λ°λͺ μΌλ‘ μλμ§λ₯Ό μ΄λ£Ήλλ€. κ° νΉνλ μ²κ΅¬νΒ·λλ©΄Β·μ νκΈ°μ λΉκ΅Β·κ΄λ ¨ λ Όλ¬Έ(References)μ ν¬ν¨ν USPTO κ°μΆμ νμ€ νμμΌλ‘ μμ±λμ΄ μμ΅λλ€. All four patents below are implemented in a real, working system (agedlearning.com) and are being validated with clinical data from Boston Neuromind LLC. DMDA (Patent 4) is the core, and the other three patents are supporting inventions that synergize around it. Each patent is written in USPTO provisional-application format, including claims, drawings, prior-art comparisons, and references.
Harvard Mind, Brain & Educationμ Kurt Fischer κ΅μκ° μ 립ν 13λ¨κ³ λμ κΈ°μ μ΄λ‘ μ μ μ©ν΄, μ¬μ©μμ μΈμ΄ 볡μ‘λ(μ΄ν·ꡬ문·결μ)λ₯Ό μ€μκ° μΆμΆνκ³ AI μλ΅ λ³΅μ‘λλ₯Ό λμ μΌλ‘ μ‘°μ νλ μκ³ λ¦¬μ¦. λ°λͺ μκ° Fischer κ΅μμκ² μ§μ μ¬μ¬λ°μ νμ μμ°μ μκ³ λ¦¬μ¦μΌλ‘ ꡬν. An algorithm that applies the 13-level Dynamic Skill Theory of Harvard's Prof. Kurt Fischer (Mind, Brain & Education): real-time extraction of a user's linguistic complexity (lexical, syntactic, cohesive features) is used to dynamically adjust AI-response complexity. The inventor β directly mentored by Prof. Fischer β implements this academic legacy as an algorithm.
19μ±λ μ λλν(qEEG)μμ μΆμΆλ λ°μ΄μ€λ§μ»€(Theta/Beta Ratio, Frontal Alpha Asymmetry, Peak Alpha Frequency λ±)λ₯Ό AI λν μμ€ν μ νλΌλ―Έν°(μλ΅ μλ, μ μ ν€, μΈμ§ λΆν)μ μ§μ λ§€ννλ μΈκ³ μ΅μ΄μ μκ³ λ¦¬μ¦. BCN(Board Certified in Neurofeedback) μ격μ κ°μ§ λ°λͺ μλ§ μ€κ³ κ°λ₯ν λ μ κΈ°μ . The first algorithm in the world to directly map biomarkers extracted from 19-channel quantitative EEG (qEEG) β including Theta/Beta Ratio, Frontal Alpha Asymmetry, and Peak Alpha Frequency β onto the parameters of an AI conversation system (response latency, affective tone, cognitive load). A proprietary technology that only an inventor holding Board Certification in Neurofeedback (BCN) can design.
μΌκ΅΄ νμ (FACS), μμ± μ΄μ¨, ν μ€νΈ λ΄μ©μ μκ°μ λΆμΌμΉ ν¨ν΄μ νμ§ν΄, μκΈ°λ³΄κ³ λ‘λ λλ¬λμ§ μλ κ°λ©΄ μ°μΈ(masked depression), μλ μν°λ―Έμ, μμ΄ μν μ νΈλ₯Ό ν¬μ°©. λμ μκ° μν(DTW)κ³Ό λ€μ€ λͺ¨λ¬ μ μ μ΅ν©μ κ²°ν©ν μκ³ λ¦¬μ¦. "λ§κ³Ό νμ κ³Ό κΈμ΄ κ°μ λ°©ν₯μ ν₯νλκ°?"λΌλ μ§λ¬Έμ κ°κ΄μ μΌλ‘ λ΅νλ€. An algorithm that detects temporal incongruence across facial expressions (FACS), vocal prosody, and textual content to capture masked depression, alexithymia, and suicide-risk signals invisible to self-report. It combines Dynamic Time Warping (DTW) with multi-modal affect fusion to objectively answer the question: "Do voice, face, and text point in the same direction?"
| νλͺ©Item | κ°μCount | κ±΄λΉ λΉμ©Per-Item Cost | μκ³Subtotal |
|---|---|---|---|
| Patent 4: DMDA β Flagship | 1 | $2,000β3,000 | $2,000β3,000 |
| Patent 1: Fischer DSL | 1 | $1,500β2,500 | $1,500β2,500 |
| Patent 2: QEEG-Conversation | 1 | $1,500β2,500 | $1,500β2,500 |
| Patent 3: Temporal Congruence | 1 | $1,500β2,500 | $1,500β2,500 |
| μ΄ USPTO Provisional μΆμ λΉμ©Total USPTO Provisional Cost | 4 | β | $6,500β10,500 |
* μ λΉμ©μ κ°μΆμ(provisional) κΈ°μ€μ΄λ©°, 1λ λ΄ μ μ(non-provisional) μ ν μ μΆκ° λΉμ© λ°μ. DMDAλ FDA 510(k) κ²½λ‘ λ³λ λΉμ© λ³λ μ°μ . * Costs above are for provisional filings; conversion to non-provisional within one year incurs additional expense. DMDA also requires a separately budgeted FDA 510(k) pathway.