update README: articulate DANA architecture

README.md

@@ -3,36 +3,45 @@
 # `OpenSSA`: Small Specialist Agents for Problem-Solving
 
 `OpenSSA` is an agentic AI framework for solving complex problems in real-world industry domains,
-overcoming the limitations of LLMs and RAG in such settings.
+overcoming the limitations of LLMs and RAG in high-precision settings.
 
-## Level-2 Intelligence with Planning, Reasoning, domain-specific Knowledge and diverse Resources
+At the heart of this framework is a __Domain-Aware Neurosymbolic Agent (DANA)__ architecture,
+which treats domain-specific knowledge as a first-class concern
+and applies captured knowledge representations in neural and symbolic program search and program execution
+to achieve consistency and accuracy in problem-solving.
 
-`OpenSSA` agents, built with powerful Hierarchical Task Planning (HTP) and Observe-Orient-Decide-Act Reasoning (OODAR),
-go far beyond the Level-1 pattern-matching intelligence performed by LLMs and RAG and achieve superior outcomes
-in complex multi-faceted, multi-step tasks. See our [comparative study](https://arxiv.org/abs/2404.11792).
+## Level-2 Intelligence with Domain-Specific Knowledge and Sophisticated Planning & Reasoning
 
-`OpenSSA` agents can also be armed with domain-specific Knowledge, connected to diverse Resources
-(files, databases, web sources, etc.), and/or be guided by specialized industry experts
-to maximize the accuracy and comprehensiveness in their planning, reasoning and deliberative/iterative problem-solving.
+`OpenSSA` implements a variant of the DANA architecture,
+with problem-solving programs represented in a Hierarchical Task Plan (HTP) form
+and program execution by powerful Observe-Orient-Decide-Act Reasoning (OODAR)
+(see [OODA comparative study](https://arxiv.org/abs/2404.11792)).
+`OpenSSA` DANA agents can also be armed with diverse Resources such as files, databases and web search.
+
+This combination of the knowledge-first DANA architecture with HTP and OODAR implementations
+goes far beyond the Level-1 pattern-matching intelligence performed by LLMs and RAG
+and achieves superior consistency and accuracy in deliberative/iterative multi-step problem-solving.
+
+## Small and Resource-Efficient Agents for Practical Real-World Deployment
+
+Such Level-2 intelligence through domain-specific knowledge and planning and reasoning
+allows `OpenSSA` DANA agents to work well in many industry applications
+using significantly smaller component models, thereby greatly economizing computing resources.
 
 ## Open and Extensible Architecture
 
 Committed to promoting and supporting open development in generative AI,
 `OpenSSA` would strive to integrate with a diverse array of LLM backends, especially open-source LLMs.
+For example, `OpenSSA` supports `Llama` LLMs and models derived or fine-tuned from them.
 If you would like certain LLMs to be supported, please suggest through a GitHub issue, or, even better, submit your PRs.
 
-Additionally, `OpenSSA`'s key Planning, Reasoning, Knowledge and Resource interfaces
+Additionally, `OpenSSA`'s core Knowledge, Planning, Reasoning and Resource interfaces
 are designed with customizability and extensibility as first-class concerns,
 in order to enable developers to effectively solve problems in their specific industries and specialized domains.
 
-## Small and Resource-Efficient Agents for Practical Real-World Deployment
-
-Specialized, Level-2 intelligence allows `OpenSSA` agents to work well in many applications
-using significantly smaller component models, thereby greatly economizing computing resources.
-
 ## Getting Started
 
-Install by __`pip install openssa`__ on Python __3.12__.
+Install by __`pip install openssa`__ on Python __3.12 or 3.13__.
 
 - For bleeding-edge latest capabilities: __`pip install https://github.com/aitomatic/openssa/archive/main.zip`__.