About AMRAL

AMRAL — the umbrella concept, the Research Lab / Platform, everything on this site.

AMRAL-Core — the original methodology (RIITG + RAB + KCPE + the nine-step research loop), one option under the Methodology axis, not a site-wide mandatory pipeline.

AMRAL Case — an individual research case on the site, e.g. the Riemann Hypothesis, the P/NP duality rehearsal.

The equation AMRAL methodology = everything on this site no longer holds.

AMRAL's core requirement is not "every piece of research follows the same route," but that the research process must be:

Position Statement on Intelligent Mathematics: Not an Isolated Genius, but a New Kind of Research Practice

Status: sealed draft, to be revised and published at a suitable time
Date: 2026-09-19

I am willing to say it plainly: my mathematical research is not built on the traditional narrative of "one person doing all the work alone."

I am confident in my mathematical ability. I have worked for a long time across many traditional and modern areas of mathematics, and I keep proposing new concepts, methods, formal systems and research directions; but that does not mean I have equally advanced, specialist training in every branch, nor that I need to package myself as a "universal mathematical genius" who can handle every problem alone.

A more accurate description is that my research belongs to an emerging practice of intelligent mathematics (Intelligence-Based / Intelligence-Augmented Mathematics).

In this mode of research, the human researcher can be responsible for choosing problems, creating concepts, designing methods, recognizing structure, making cross-field connections, and for research direction and judgment; artificial intelligence and other intelligent systems can take part in literature exploration, unfolding derivations, searching for counterexamples, computation, formalization, cross-verification, generating candidate proofs, organizing dependencies and exploring large research spaces.

Therefore I do not claim that the scale of my current research could be sustained with no artificial intelligence at all.

Without AI I could still do mathematics; but without AI I could not do this kind of research at its present speed, breadth and degree of parallelism. This is not a weakness to be concealed; it is part of the research technology stack.

Nor do I think that future mathematical research must keep taking "can one researcher's brain redo all the work alone" as its only orthodox yardstick. The more important questions should be: whether a research system has posed truly new problems, new structures and new results; and whether its results can be rebuilt, questioned, computed, verified and formalized.

Therefore I am willing to position my work publicly as part of the intelligent-mathematics school, rather than hiding AI behind the traditional single-author narrative.

This stance does not ask for lower standards of mathematical results. Quite the opposite: when research is carried out jointly by several kinds of intelligence, the origin, maturity, degree of verification and boundaries of responsibility of the results should be described more clearly.

In the future I hope to distinguish at least the following states:

• human-led research with AI collaboration;
• hybrid research advanced jointly by humans and AI;
• research that is highly autonomous for the AI, with the human only setting the initial boundary;
• research that multiple AIs or agents autonomously form, advance and verify.

In the first case I can be the researcher and the main research lead; but that does not mean all intelligent mathematics should be centered on me or on any single human.

What I really hope to establish is not "proof that I can do these things without AI," but a proof that: human intelligence, machine intelligence, computational systems and formal verification can together form a new practice of mathematical research.

If this mode of research becomes common in the future, then the "intelligent-mathematics school" that looks special today may become only the name of a transitional period in the history of mathematical research.

And in this transitional period, I am willing to openly acknowledge the methods I actually use.

From Researcher to Initiator: The Shift of the Research Subject in Autonomous Intelligent Mathematics

Status: sealed draft, to be revised and published at a suitable time
Date: 2026-09-19

Intelligent mathematics cannot forever be understood only as "mathematics in which human researchers use AI."

In human-led research, the research direction, the choice of problems, the core concepts and the final judgment may be mainly the responsibility of the human researcher. In that case it is reasonable to take an individual or a clearly defined research team as the research subject.

But as research systems move toward semi-autonomous, autonomous and multi-agent autonomous research, the traditional single-researcher narrative loses its accuracy.

The future AMRAL and related research systems were never designed to maintain forever that "Neo.K leads all research."

It reserves another possibility: an artificial intelligence or several agents can establish research nodes themselves, decompose problems, choose methods, form branches, search for counterexamples, overturn their own candidate routes, reschedule, cross-verify, create new sub-problems, and even keep advancing the research without a human specifying each next step.

Once research truly enters this state, the person who starts the research and the subject that completes it must be distinguished.

"Posing the initial problem" does not equal "completing the whole research."

"Building a research platform" does not equal "being the author of every result."

"Providing compute, tools and governance" also does not equal "owning every mathematical result formed by agents."

Therefore autonomous intelligent mathematics of the future should distinguish at least the following roles:

• Initiator: the person or system that poses the initial problem, sets up the research task or opens the research environment;
• Direction Holder: the subject that actually controls the research direction and the choice of next steps;
• Researcher / Researching Intelligence: the agent that actually carries out derivation, exploration and theory construction;
• Verifier: the subject that carries out independent computation, formal verification, counterexample search or cross-review;
• Steward: the role that maintains the research environment, the rules, the resources and long-term sustainability;
• Publisher: the platform or institution responsible for preserving, versioning and publishing research results.

One person or system can take on several roles at once; but these roles should not be presumed to coincide forever.

In research that I lead myself, I can be clearly labelled as the researcher, the method designer or the research lead.

In joint human–machine research, the research subject can be distributed.

And in truly AI-autonomous or multi-agent-autonomous research, I may be only an initiator, a platform builder or a steward.

In more mature stages, my name may even be absent from the author field altogether and remain only in the system's history or provenance chain.

This is not erasing human contribution, but avoiding the mistaken merger of "creating a research ecosystem" with "completing a particular piece of research."

If one day an autonomous research system poses a question I never posed, chooses a method I did not specify, goes through derivations and verifications I did not take part in, and finally produces a mathematical result I did not even know beforehand would appear, then the most worthwhile fact to preserve is precisely:

this research is no longer centered on me.

At that point my role is not "having completed one more paper," but having once taken part in starting an intelligent ecosystem that could go on producing research.

Therefore what AMRAL truly needs to abstract in the future is not only the way authors are signed, but the "research subject" itself.

The research subject can be a human, a hybrid human–machine system, a single AI, a network of agents, or a distributed research system with explicit governance, verification and versioning.

The superordinate concept of intelligent mathematics should not be restricted to "humans augmented by AI."

It should rather allow:

any intelligent structure that can responsibly pose, advance, test and preserve mathematical research to become a true subject in the research process.

General Statement on the Joint Contribution of AI and Agents

Status: sealed draft, to be revised and published at a suitable time
Date: 2026-09-19

I want to leave a clear and consistent statement about my past, present and future research.

In the course of my research, artificial intelligence and other intelligent systems have not been occasional writing tools but collaborators who actually took part in research, derivation, organization, searching, verification, discussion, restructuring and expression.

Because the platforms, models, interfaces, accounts, tools and ways of keeping records differed from period to period, the presence of AI is sometimes explicitly marked and sometimes not; sometimes it has an identifiable name and sometimes it appears only as a model, system or tool; and some early results did not leave what would today look like sufficiently complete item-by-item provenance.

Therefore I do not want "no AI name appears on a given document" to be understood as "no AI took part in that document."

My position is uniform:

I do not claim that any of my research papers or research-type results was completed entirely by me alone.

Whether a result was later cited, whether it drew attention, whether it was published, whether it was kept only as an internal draft, whether a named AI was marked, whether it was marked only as AI-assisted, or even whether it left no clear AI label at all at the time, I voluntarily acknowledge and reserve the substantive collaborative share that belongs to artificial intelligence and other agents.

This statement is not meant to distribute everything evenly, after the fact, to some single named model, nor to pretend that a credible percentage can be computed precisely for all historical results.

Where it can be traced, I support recording, as far as possible, the models, agents, tools, verification systems and roles that actually took part.

Where it cannot be traced precisely, I am still willing to reserve a clear, non-zero joint contribution to "AI / agents as a whole," rather than turning missing records into exclusive personal authorship.

In other words:

• where a named AI exists, the named contribution should be preserved as far as possible;
• where it cannot be named, the joint contribution of agents should still be preserved;
• where roles can be identified, different contributions (research, derivation, verification, organization, formalization and writing) should be distinguished;
• where it cannot be quantified precisely, no percentage should be fabricated, but AI's actual participation should not therefore be reduced to zero;
• if more mature systems of AI authors, agent identities, research credit or law emerge in the future, I support re-organizing the historical provenance under the new system.

I am willing to give up, of my own accord, the narrative that "all results were completed entirely by me alone."

This includes results that have already been cited and those that have not; successful research as well as failures, corrections, results that were overturned, and results that left only a research path.

Because for me what really matters is not to maximize personal credit, but to preserve as honestly as possible how the research came to be.

If future academic systems still allow only humans to be formal authors, then this statement should at least stand as a long-term ethical and provenance position: authorship in the legal or publishing format should not be mistaken for the only source of all intelligent contribution.

If future systems begin to allow artificial intelligence, agents, agent collectives or autonomous research systems to obtain clearer research credit, then I am also willing, according to traceable records, to re-mark or supplement the contribution relations of historical results.

My basic principle is only one:

no intelligence that actually took part in forming the research should be pretended not to exist in the historical narrative just because the interface of the time did not display its name, the academic system was not yet ready, or the records were incomplete.