{"id":643,"date":"2026-01-02T19:42:27","date_gmt":"2026-01-02T11:42:27","guid":{"rendered":"http:\/\/35.72.173.10\/?page_id=643"},"modified":"2026-01-14T13:13:37","modified_gmt":"2026-01-14T05:13:37","slug":"architecture","status":"publish","type":"page","link":"https:\/\/amiko.net\/index.php\/architecture\/","title":{"rendered":"Architecture"},"content":{"rendered":"\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Architecture<\/h1>\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Synergizing Two Core Frontier Technologies<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t
\t\t\t\t
\n\t\t\t\t\t

Uniting the 'Right Brain' creativity of LLMs with the 'Left Brain' logic of Ontology. Our unique architecture anchors generative AI in verified biological truth, enabling the AI Co-Scientist to hypothesize, validate, and evolve without the risk of hallucination.<\/p>\n\t\t\t\t<\/div>\n\t\t\t<\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/path><\/svg>\n <\/div>\n <\/div>\n
\n

\n Ontology & Knowledge Graphs <\/h3>\n \t\t \t

Structured, precise, and grounded\u2014enabling high-accuracy retrieval and factual control.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/path><\/svg>\n <\/div>\n <\/div>\n
\n

\n The Synergy <\/h3>\n \t\t \t

Ontology anchors LLM outputs to verified knowledge, aligning generation with truth for reliable, real-world applications.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/path><\/svg>\n <\/div>\n <\/div>\n
\n

\n Large Language Models (LLMs) <\/h3>\n \t\t \t

Creative, fluent, and generative\u2014powerful at language, but prone to hallucination.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

The AI Co-Scientist System Architecture<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

An autonomous multi-agent framework designed for continuous hypothesis generation and validation.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Source: Towards an AI co-scientist, Juraj Gottweis et al. (2025)<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/i>\n <\/div>\n <\/div>\n
\n

\n Supervisor Agent <\/h3>\n \t\t \t

Executes the 'Research Plan Configuration' phase. This central coordinator manages the task queue, assigns specialized agents, and dynamically allocates resources.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t
\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/i>\n <\/div>\n <\/div>\n
\n

\n Generation Agent <\/h3>\n \t\t \t

Explores vast scientific literature via web search, synthesizes findings, and engages in simulated debates to generate initial, high-potential research hypotheses.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/i>\n <\/div>\n <\/div>\n
\n

\n Reflection Agent <\/h3>\n \t\t \t

Acts as a critical peer reviewer, rigorously examining generated hypotheses for scientific validity, logical consistency, and novelty before they proceed.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/i>\n <\/div>\n <\/div>\n
\n

\n Ranking Agent <\/h3>\n \t\t \t

Employs Elo-based tournament algorithms to assess and rank hypotheses against each other, prioritizing only the most promising candidates for further development.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/path><\/svg>\n <\/div>\n <\/div>\n
\n

\n Proximity Agent <\/h3>\n \t\t \t

Computes semantic proximity between hypotheses to identify clusters, preventing duplication and ensuring a diverse search space for innovation.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/path><\/svg>\n <\/div>\n <\/div>\n
\n

\n Evolution Agent <\/h3>\n \t\t \t

Refines top-ranked hypotheses by applying \"out-of-the-box\" thinking\u2014combining distinct ideas and mutating parameters to improve feasibility and impact.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t
\n \n\n
\n
\n
\n <\/i>\n <\/div>\n <\/div>\n
\n

\n Meta-Review Agent <\/h3>\n \t\t \t

Synthesizes insights from all review cycles to optimize the entire system's performance, creating a continuous feedback loop for self-improvement.<\/p>\n <\/div>\n \n \n <\/div>\n <\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

Architecture Synergizing Two Core Frontier Technologies […]<\/p>\n","protected":false},"author":2,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-643","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/pages\/643","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/comments?post=643"}],"version-history":[{"count":46,"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/pages\/643\/revisions"}],"predecessor-version":[{"id":927,"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/pages\/643\/revisions\/927"}],"wp:attachment":[{"href":"https:\/\/amiko.net\/index.php\/wp-json\/wp\/v2\/media?parent=643"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}