Artificial Intelligence T 2 Overview Artificial Intelligence AI
- Slides: 23
Artificial Intelligence & T 2
Overview Artificial Intelligence (AI) Goals for Applying AI to Patents Patent-Specific AI Center for AI & Patent Analysis at CMU 2
AI Artificial Intelligence (AI) Software that performs activities we consider to require human intelligence Reasoning, problem solving, learning 3
Identifying objects in an image or video Examples of AI Tasks 4
Play complex strategy games at a human level Examples of AI Tasks 5
Natural Language Processing (NLP) Natural Language Processing Subfield of AI that processes ‘natural’ languages (e. g. , English, Spanish) Process text in a document 6
Apple Siri, Amazon Alexa NLP Examples NLP software that interprets spoken commands Spam Filters NLP software that identifies likely spam emails Patentability Searching NLP software that finds “similar” patents 7
Natural Language Processing (NLP) NLP can find and extract desired information from massive amounts of text Reduces the need for humans to read as much 8
Goals for Applying AI to Patents AI is not magic In fact, AI doesn’t have much “intelligence” AI cannot understand text like a human We must set our expectations accordingly 9
Goals for Applying AI to Patents High Level AI Goals: AI helps the user read less, search less Locate desired information within patents Summarize information A collaboration between human and AI Each doing what they’re best at 10
Goals for Applying AI to Patents High Level AI Goal: The better the AI, the less the human needs to read 11
Goals for Applying AI to Patents More Specific AI Goals: “Better” AI understands exactly what information the human needs for a particular task This allows the AI to provide only the most relevant information to the user Reduce reading time 12
Goals for Applying AI to Patents Example: Freedom to Operate “Better” AI can help more if it understands: Patent claims What the claims mean How to assess claim infringement Whether a particular product might infringe Patentability requirements of claims Whether particular claims satisfy patentability requirements 13
There are many tasks specific to the patent domain: Patent. Specific AI Reading and understanding patent specifications Reading and understanding claims Finding potential infringers of claims Finding potential licensees Assessing patent validity Conducting freedom to operate analysis Conducting white space analysis Analyzing patent citations 14
Patent. Specific AI to help the user read patent specifications & claims Auto-summarize i. e. create a brief, meaningful summary 15
Patent. Specific AI to find potential infringers or licensees 1. Search for product descriptions, and then 2. Apply claim infringement analysis Compare product to each claim in portfolio 16
Patent. Specific AI to assess patent validity For each claim, evaluate validity for Anticipation Obviousness Indefiniteness Lack of written description 17
Patent. Specific AI to perform freedom-to-operate analysis 1. Analyze a product description, and then 2. Apply claim infringement analysis Compare product to third party claims 3. Apply validity analysis to infringed claims 18
Patent. Specific AI to perform patent whitespace analysis 1. Evaluate a portfolio of claims 2. Determine scopes of claims 3. Identify areas where claim coverage is sparse or nonexistent 19
Patent. Specific AI to analyze (improved) patent citation networks Reduce noise in patent citations Use citations that reflect ‘true’ patent similarity 20
Center for AI & Patent Analysis (CAPA) at Carnegie Mellon University CAPA’s Mission To create patent-specific AI AI algorithms and tools for work done by Patent attorneys Licensing professionals Corporate managers Patent examiners Many short- and long-term projects 21
I invite everyone to discuss their workflow and challenges with me Conclusion Help us design AI for the most common tasks I may already have software that can facilitate your work You may want to join an existing cosponsored project in process 22
Layer 4: Legal Analysis Compare scopes of claims Summarize Office Action issues Assess claim rejections and objections Compare claims to product descriptions Search for patents infringed by a product Layer 3: Patent Knowledge Appendix: Patent Model Initiative Basic claim limitations Basic claim interpretation alternatives Possible ambiguity Possible gaps in claim meaning Possible gaps in claim coverage Structure for means-plus-function limitations Hierarchy of terms used in patents Layer 2: Structured Patent Information Patent and Application data Claim components, elements, steps Terms Relations among terms Definitions Embodiments and examples Boilerplate language Characterizations of prior art PTAB & Court opinion data Claims and limitations Arguments and rebuttals Statements on evidence References Cited authority Conclusions and orders Office Action data Claim rejections & bases Claim amendments Interpretations of claims Characterizations of art Proffered evidence Arguments & inferences Cited authority Responses and rebuttals Layer 1: Raw Text Patents and applications Office Actions and responses PTAB & Court opinions 23
- Business intelligence overview
- Areanna ai
- Conclusion of artificial intelligence
- Steps towards artificial intelligence
- Inference by enumeration in artificial intelligence
- Artificial intelligence devices
- Ai is a branch of computer science
- Uas kecerdasan buatan
- Optimal decisions in games in artificial intelligence
- Unit 7 artificial intelligence
- Rote learning in artificial intelligence
- Int404
- Iterative deepening search prolog
- Differentiate between strips and adl
- Fuzzy proposition
- Artificial intelligence chapter 1
- Part picking robot peas
- What is utility based agent
- Blind search algorithm
- Artificial intelligence
- Computational intelligence ppt
- Analogizers
- Athena machine learning
- Uninformed search in artificial intelligence