Synergy Engine

Synergy Engine

The Synergy Engine is a distributed search and query engine over the entire ADAM ecosystem, which means its queries also work over any underlying technology protocol. The Synergy Engine leverages the collective intelligence of agents in the network, is owned and run by everyone, and does not rely on any central servers or opaque algorithms and AI.

The ADAM layer represents a significant shift from a centralized model to a distributed, agent-centric paradigm. This transition creates a new kind of web that is interoperable, semantic, and based on human communication. However, the decentralized nature of the ADAM ecosystem raises the challenge of finding data within the network, as data is distributed among individual agents. This challenge is closely connected to the broader concept of sense-making, where finding reliable information involves understanding sources, evaluating trustworthiness, and identifying semantic connections between data.

To address these challenges, the Synergy Engine enables the query of information by considering intrinsic properties of the requested data and evaluating the trustworthiness of its source.

How it Works

• An agent can send a query through the network and ask for specific information

• Agents in the ADAM ecosystem can propagate queries by checking their own perspectives for matching data

• The ADAM layer uses Social DNA, which consists of Prolog programs, to define graph patterns in ADAM, and Social DNA is used in queries to validate the relevance of a result

• Social DNA allows nodes in the network to verify results independently and decentralizes trust

• Agents that provide query results, or pass a query to agents in their social graph that might have the desired results, are rewarded in the SynergyFuel needed to make a query

• The Synergy Engine harnesses collective intelligence, rather than centralized and artificial intelligence, to find and validate information

The Social Stack

• Queries in the ADAM network trace a path through multiple agents, creating a "Social Stack" that records the agents the query has passed through

• The Social Stack consists of ADAM links that describe the relationship between the previous agent who forwarded the query and the next agent who received it

• Each ADAM link is signed by the previous agent, providing cryptographic proof of their interaction and relationship

• If an agent cannot resolve a query, they can pass it on to their friends by creating an ADAM Link with themselves as the source and the receiving agent as the target

• The Social Stack is included in query results and must show a connection between the query source and the agent providing the result

• Queries can specify conditions relating to the Social Stack, such as requiring a certain level of trust between agents in the stack

• The control over the Social Stack allows users to define complex requirements for valid query results

Searching Via Other People’s Social Graphs

ADAM's distributed search and query process leverages social connections and social graphs to enable networked exploration. It uncovers "social cliques" of interconnected agents with similar perspectives, offering different query results. These cliques reflect the diversity of perspectives in a distributed network, acknowledging the contextual nature of knowledge and understanding.

Coasys aims to build infrastructure to connect with agents outside one's social circle and create global public indexes. This approach facilitates comprehensive understanding and fosters collaborative knowledge generation in an inclusive manner, offering a nuanced way to navigate the information ecosystem.

Privacy Concern and Solution

ADAM's query mechanism, based on agents sharing information through interconnected networks, raises privacy concerns regarding the potential sharing of confidential data without consent. To address this, ADAM implements a systemic solution by including watermarks in every expression, preserving the integrity of private information and representing the context and visibility rights. Sharing data with a watermark in query results would break trust publicly, leading to downgraded trust and exclusion from social networks. Data ownership is tied to the social organism with the correct watermark, ensuring privacy and maintaining the network's credibility while respecting data privacy within the ADAM ecosystem.

Relationship to AI and LLMs

ADAM combines artificial and collective intelligence in a thriving ecosystem of applications where users can incorporate their privately held perspectives into query results. The principle of collective intelligence challenges centralized AI models by offering more authentic and trustworthy outcomes. Instead, AI and LLMs are used mainly translate a user's query input, and human trust relationships and distributed data sources are used to provide reliable and authentic information.

The Synergy Engine allows users to interact with ADAM like a centralized AI model, but with responses grounded in the reality of collective intelligence and human trust relationships, integrating AI as a powerful tool for an enhanced user experience.