A yet2 client seeks start-ups and innovative companies developing novel artificial intelligence, machine learning (AI/ML), and advanced computing technologies, products, and platforms to aid in the development, training, and deployment of AI-based solutions across the organization’s broad business areas and offerings. This organization is interested in AI/ML technologies (and associated technology providers) in four broad technical categories (listed below), and is particularly interested in technologies and systems within these categories designed for high-stakes applications and use cases.
Categories of Interest
The client is interested in identifying and establishing an ecosystem of AI/ML companies and tools across four broad topic areas of interest:
- Context sensitive AI/Behavioral Flexibility – given the trend of increasing autonomy of machine learning and machine reasoning architectures, and ability of these systems to synthesize learned and symbolic knowledge, highly context-sensitive and flexible systems become both possible and necessary for many high-stakes applications.
- Including, but not limited to: Neuro-Symbolic Architectures – integration of ML and RL, Hierarchical Reinforcement Learning, Affordance Learning, Multi-modal object classification in visual and Electromagnetic Spectrum in high ambiguity context
- On-line inference/decision making – deliberative consideration/analysis of various possible solutions/courses of action during high-stakes, “Mission critical”, and/or real-time situations
- Including, but not limited to: Mental simulation/approximate simulation/accelerated on-line simulation, Advanced Optimization Architectures using Optimization techniques
- Computational Orders – systems that can interpret natural language orders and automatically validate and audit actions associated with subsequent orders and plans generated by humans or machines are of interest
- Including, but not limited to: Cognitive Engineering/Human Machine Symbiosis, Ontologies/Ontology building methods, OpenAI’s GPT-3 Model, Computational Law
- Explainable AI (XAI) and AI Verification and Validation (V&V) – solutions that build trust in systems containing AI and their associated actions and outputs through model interpretation, model completeness understanding, bias identification, and other transparency methodologies
Possible Solution Areas
The client is open to solutions from any industry or application; however, they are particularly interested in solutions being developed or currently used in applications in high-stakes industries such as financial services, gambling, security, defense, public safety, autonomous vehicles, aerospace, fraud detection, predictive maintenance, and more.
Requirements / Constraints
Solutions of particular interest will have the following characteristics:
- Be able to scale or currently used for large-scale production applications
- Deployed on-prem and/or at the edge
- Kubernetes-based solutions
- Open-source solutions are of interest, but not required
- Early-stage companies (e.g. pre-revenue or just begun generating revenue) are of particular interest
University research is not of interest.
Desired Outcome of the Solution
yet2’s client is open to a range of collaboration opportunities
Related Tech Needs