Research and Learning

Most of the time, attend to the whole; some of the time, develop familiarity with important points and details, and know when to examine them. Begin with a field's questions, concepts, and relationships, then deepen particular parts as needed. Prioritize broadly applicable understanding and methods.

Design is one example: deciding what to make, why, how to organize it, and what would count as a fitting result. When AI carries implementation, people can concentrate their learning on these questions. Designing twice or more means generating and comparing alternatives, reconsidering the problem, premises, and evaluation criteria when needed. Even the framework organizing practice and reflection can change; this is the sense of reasoning beyond the current double loop used here.

First distinguish what you want to learn

Theory and practice support each other; unless the current purpose is purely theoretical inquiry, begin with practice. Identify what you want to accomplish: the abilities an employment direction needs, the problem a project addresses, or what an assignment asks you to explain or produce. Work backward to the relevant concepts, principles, and methods.

This is bidirectional pruning: practical demand selects the scope and depth of theory, while theoretical dependencies identify missing foundations. When the selected theory remains unclear, supply what makes it intelligible and return to the original question. Theory can also expose mistaken practical assumptions and unnecessary steps. Adjust scope to actual difficulties. Pure theoretical inquiry can follow its own questions, concepts, and arguments.

If you want to give yourself a university education, or fill important gaps in foundational and practical knowledge, the five steps below provide a systematic route: draft a curriculum from your interests, prepare courses, books, tools, and an environment, then write theory and practice paths and begin learning. You can cover one important field or gradually develop your own curriculum.

Another entrance begins with books or papers. Becoming familiar with an unfamiliar field through these alone can be harder to organize: papers often omit shared background, and a book may assume foundations from other disciplines. Models can help reconstruct the missing knowledge relationships, while the purpose of reading determines how far to pursue them.

These entrances can connect. Research reading can reveal a substantial gap that calls for a course, while everyday reading can uncover a question worth pursuing. The materials, environment, and path preparation below belongs to the systematic course route; materials for a single paper can develop during reading. Either teacher can support these entrances: choose the convenient teacher for the current activity.

Reference materials and teachers

For a systematic course, a book, a single paper, or a question pursued across several papers, give the source and learning purpose to Dot or local Codex. Have the model use a Teacher Skill to write connected reference materials, then work with a teacher to read, explain, and discuss them. Preparing materials and using a Teacher support all these learning entrances.

This self-study route starts from native-language texts and AI-prepared material, without assigning original lectures to watch or listen to. In the systematic course route, read the main textbook, chapters, or papers used by the selected course in the learner's native language, or in translation, before that course's AI preparation files. Books already written in the learner's native language can be read directly; have the model translate foreign-language sources first. Original texts and faithful translations have a different role from AI-written explanations, guides, and practice. See the reading priority among materials.

Depending on the learning object, use the two- or three-document arrangement below. The model explains concepts and relationships, retains chapter, lecture, page, or paper-passage locators, adds examples and useful diagrams, and connects existing understanding, interests, and actual questions:

Organize materials around one learning object. A single paper can have its own two or three reference documents; papers addressing one question can share a set while retaining the relevant locations in each source. Overall learning paths select across courses; a specific guide develops understanding of the current book, paper, or question. Original sources and model explanations retain their separate identities.

Preparation depth follows purpose. A systematic course can have its full materials prepared before study begins. Single-paper reading and research learning can start with the materials currently needed, then add foundational papers' knowledge and relationships to the glossary and guide as gaps emerge. Everyday learning can use the same arrangement, emphasizing important abstractions, conclusions, and examples, with a few understanding checks where helpful.

Choose a convenient teacher

Both share teaching methods; begin with the entrance convenient for you. Ask directly: “Guide me through this paper,” “Explain how this concept relates to my existing understanding,” or “Help me explore this research idea.” Teachers can also use the disciplinary references and research methods below. Let the model handle downloading, installation, connections, and selection. See Using the Tools for general tools and file arrangements.

I. Build your own digital university

Organize systematic study of foundational and practical knowledge through the five steps below. Return between steps when preparation calls for revision; once ready, theoretical and practical learning can develop together.

1. Draft a curriculum from your interests

Explore interests and compare curricula

Begin with an interesting direction, such as human factors. Discuss with AI whether you want to understand cognition, human–machine cooperation, or the design of products and working environments. Concrete questions can gradually clarify the direction.

Ask AI to investigate Chinese universities with established work in that direction and find recent master's curricula and entrance-examination reading lists for the relevant programs. Related material may appear under ergonomics, engineering psychology, or human–computer interaction; judge the actual research groups and courses. Extract courses, topics, prerequisites, and books, retaining the university, department, year, and original link.

Where programs emphasize engineering, psychology, or design, or sit in computer science, psychology, artificial intelligence, design, or other departments, select representative curricula. Have AI explain the questions, knowledge, and methods each emphasizes.

Draft a course system around one main curriculum

Ask AI to judge which curriculum fits your interests, responses to courses, and intended activities, explaining its reasons so you can choose or revise the judgment. Use the most interesting curriculum as the main structure and fill gaps from the others: merge duplicates, add missing foundations and methods, arrange prerequisites, and distinguish core courses from interest-led extensions. Each course should have a clear purpose and relationship to the others.

The result at this point is an adjustable curriculum draft. As materials and tools are prepared, record which courses, books, and practical abilities can connect; organize the learning paths once those resources are in place. Preparation may reveal an uninteresting course, unsuitable materials, or changed tool conditions. Return to replace a course, revise the curriculum, or choose a more interesting direction when needed.

2. Turn courses and books into readable materials

Find matching resources and download them

Search primarily for online courses from outside China with matching topics or course names, selecting the latest available versions. Check syllabi and teaching years; explain content matches across different names and identify older versions when those are all that is available. Use the earlier entrance-examination reading lists and curricula to select and assess courses. Once a self-study course is selected, follow its syllabus, textbook list, and session plan to obtain the main textbook, chapters, and papers it actually uses, identifying their corresponding units.

Have the model first obtain the main textbook, required chapters, and papers assigned or actually used by the selected course, then the relevant lecture notes, slides, subtitles or transcripts, assignments, and code. Organize them by course and check completeness. Online courses supply a curriculum and source materials here, not lectures the learner must watch or listen through. Have the model process video only when necessary to obtain relevant subtitles or figures. Where the model cannot complete a download, list the material, original link, and reason for manual completion. Distinguish downloaded materials, online-only entrances, and outstanding items in the index.

Translate the sources before writing preparation files

Here, the original books and papers belong to the self-study course already selected: its main textbook, chapters assigned for a unit, or papers actually discussed in the course. Establish that correspondence before preparing readable editions; prioritizing original works does not call for assembling a separate extracurricular reading list. Where the course uses no corresponding textbook or paper, use the AI preparation files built from its existing sources.

When a source's language is unsuitable for direct reading, have the model translate the complete book, chapter, or paper selected for study into the learner's native language. For the Chinese learner described here, prepare a Chinese translation. Preserve chapter, page, or paragraph correspondence, equations, figures, quotations, and references; check terminology and consequential passages. Use an appropriate existing Chinese edition when available, and read books originally written in Chinese directly. Translation makes the original work readable; a summary does not replace the book or paper.

Then use these sources to write the two or three preparation files described above: a glossary, main guide, and relevant exercises or experiments. These files help explain structure, clarify concepts, supply background, and select practice. Where the original already develops the subject clearly, the guide points to its passages and adds the necessary explanation; finishing the guide is not a prerequisite for reading the original.

Reading priority among the materials

Main textbook, chapters, or papers used by the selected course, in the learner's native language or in translation → that course's two or three AI preparation files → its necessary lecture notes → slides → transcripts.

This is a priority for selecting reading, not a requirement to complete every layer. Use a sufficient textbook or paper adopted by that course as the main text; where the course uses no corresponding original, begin with its existing AI preparation files. Lecture notes, slides, and transcripts are reserve sources, translated and organized by the model for the relevant need, rather than an additional full set of unprocessed course material to read. Retain brief course or unit overviews for orientation, without replacing the main texts or returning the learner to a lecture-watching sequence.

Check the preparation files and reading entrances

Check that the original or translated texts and the two or three preparation files open correctly and link to the corresponding passages. Where training materials are useful, prepare questions, solutions, and evaluation grounds across the course's scope, checking answers and keeping them separate from questions for later selection. This systematic course route prepares materials, an environment, and learning paths before reading and practice; other reading entrances can develop materials during reading.

3. Prepare practical tools and an environment

After preparing materials and notes, have the model connect the environment needed by the curriculum draft, acquired resources, and abilities you want to develop. Software may be installed for later use or operated with model assistance. Installation, payment, usable capability, and personal familiarity are separate judgments.

Have the model inspect existing applications, plugins, and public interfaces, select a combination that serves the current work, and verify it with a small real operation: read the material, perform the action, save the result, reopen it, and continue.

For paid tools, first understand licensing, subscriptions, and feature limits; purchases follow the person's actual authorization. When retaining a fixed version matters, record the usable version, disable its automatic updates, and preserve recoverable installation materials. Select versions for actual compatibility and use.

Retain the tool capabilities, usable materials, and operation entrances established during preparation for the next step of writing the paths. Once practice begins, return what the tool is useful for, how to express requirements, how to assess results, and what you want to learn next to the practice path. The model can maintain software inventories and installation records; the human reading entrance should explain possible activities, their course connections, and how to begin. Model-facing collaboration accounts and personal materials may support this arrangement; the readable path explains available activities, while personal records stay private.

4. Write the theory and practice learning paths

Once the materials and practical environment are ready, have the model organize the connections developed during preparation into two usable paths. The theory path should point to acquired courses, books, notes, and passages; the practice path should point to available tools, materials, operation entrances, and abilities to develop. The earlier curriculum draft supplies direction; this step supplies grounded, specific learning paths.

The theory path organizes knowledge; the practice path organizes abilities and tools. A course usually includes both theory and practice, but these forms of organization answer different questions. Courses help explain concepts, principles, and methods; practice develops the abilities to design, analyze, and make things. Difficulties encountered in practice reveal which theories deserve closer study.

Write the practice path as an outline of abilities and tools. Design might distinguish paper sketches, visual expression, information and interaction structure, prototypes, motion, platform differences, and design handoff. Statistics might distinguish data preparation, visualization, model selection, interpretation, and checking. These are capability areas; the particular work, dataset, or task used for practice can be chosen when learning actually takes place.

For employment, have the model examine real job requirements and work materials, separating common foundations, specialties, experience requirements, and alternative tools. For curiosity, creation, or research, select from the relevant works, questions, and methods.

Keep an existing curriculum when it fits, adding the practice path and its connections to courses. Paths support selection and continuity. Check that both paths lead to the corresponding materials and usable environment; return any gaps to preparation, then begin reading and practice. Interests and difficulties encountered during learning can reopen the paths and their preparation.

5. Read, practice, and examine important points

Begin actual study after preparing the materials, environment, and paths. During preparation the model organizes and checks resources and tests tools. Once everything is ready, open it when you feel good and want to study.

Choose the convenient teacher through the shared teacher entrance, then begin with the prepared materials.

Begin with the current question

Follow the reading priority: open the main textbook, chapter, paper, or Chinese translation used in the selected course's current unit. A brief unit overview can orient the reading before following the author's development. Use the prepared glossary and main guide when terminology, background, or an argument is difficult, then return to the corresponding original passage. Where that course uses no corresponding original, begin with its AI main guide. Have the model check the cited original when a translation, derivation, or piece of evidence needs verification, and correct the materials accordingly.

Use diagrams to see relationships and discussion to express your understanding and questions. Ask the teacher to connect new and existing knowledge, compare explanations, and develop what remains unclear, then check whether it misunderstood. Saved notes and current explanations can revise one another: add useful examples, explanations, and new questions to the existing materials.

Work backward from practice to theory, then return

Select an ability currently needed from the practice outline. Have the teacher use the actual object to specify what to do, compare, and check. For example, begin by comparing interface transitions; when a response makes the system's state hard to judge, enter the relevant theory of motion, feedback, or perception. When a dataset needs interpretation, select an appropriate statistical model and its conditions. If the theory remains unclear, supply the foundations it actually depends on, then bring that understanding back to the original design or analysis.

The path supplies an outline. Concrete tasks, theoretical passages, and the question to return to are selected during the activity, with sequence determined by actual knowledge dependencies. Choose and change practice objects as needs develop. The necessary theory may be small or warrant substantial depth, depending on the problem and the remaining gap in understanding.

Tool familiarity includes knowing what a tool can do, how files and operations are organized, when to use it, how to express an intention, and how to judge the result. Models can operate complex interfaces, write code, and carry repetitive production; people can participate through paper sketches, photographs, speech, comparison, and revision. Preserve hands-on attempts when developing your own drawing or direct-operation ability. Where a job or examination requires personal performance, check against that requirement. Report what you performed and what the model supplied when evaluating.

When to examine important points and details

Return to relevant concepts, conditions, and evidence when a detail could change the overall judgment, results depart from expectations, understanding becomes inconsistent, or a method is entering a new context. Identify the specific difficulty with the teacher, then use examples, counterexamples, reasoning, or experiments to examine it. For example, moving from correlation to a causal explanation calls for identifying the additional assumptions and whether evidence supports them.

Actual practice should exclude most prepared material, retaining only the very few points essential now. Add diagrams, concrete examples, and smaller exercises at those points; repeat or change contexts when needed until understanding supports the current judgment, then return to the whole. Unselected materials remain available for later needs or review after forgetting.

Practice develops understanding; testing supplies performance evidence; evaluation uses that evidence to decide what needs improvement. Explanations, responses to counterexamples, and practical work can also supply evidence. When judging personal understanding, distinguish your contribution from model assistance. Assess mastery through explanation, application, and correction.

Research also requires checking the result

Learning can serve curiosity and understanding. Research develops a reasoned, testable contribution to an unresolved question and includes continued learning. Personal understanding and the reliability of research results need separate judgments. Have the model check citations and claims against original sources, examine premises, calculations, and code, and cross-check through other methods, counterexamples, or reproduction. People need to understand the consequential assumptions, evidence, and limitations to judge whether the result answers the question. Return to the relevant course when a knowledge gap appears.

II. Learn an unfamiliar field through papers

When a paper is inaccessible, find an understandable starting point through its references

Choose a target paper relevant to your question. Ask the model to explain its problem, contribution, and the concepts or inferences blocking understanding. When background knowledge is missing, examine that gap. Have the model inspect the references and search backward for more foundational papers explaining those concepts, results, or methods. Inspect the material to judge which source supplies the missing foundation. A textbook chapter or review can provide a bridge where needed.

Begin with material you can connect to what you already understand. Grasp its question, main conclusions, conditions, and the basic argument or proof supporting them. Return to the paper that uses it and explain how this newly understood knowledge supports the next step. If a gap remains, look further back; when the current dependency is understood sufficiently, work forward toward the target. The model can diagram these relationships and retain original passages so you can see why each paper is being read and what it contributes.

Proceed through the needed concepts until the original target paper becomes understandable, then move to another related paper. Neighboring work introduces new concepts, different methods, and limits to previous conclusions. Repeat the process for the new knowledge that matters. Understanding the necessary surrounding literature means covering the key dependencies, representative methods, and important disputes for the current question well enough to understand and compare them. Expand when a new question requires it.

Move from important questions toward methods that interest you

The model can also help assess the importance and feasibility of the target paper and related research directions. Why does its problem matter? What does it contribute beyond existing work, and what remains unresolved? What knowledge, data, equipment, time, or collaboration would you need to participate? Tie these judgments to papers, reviews, and actual resources, separating established facts from conjecture.

Importance and interest can guide different reading choices. An important paper may address a worthwhile question through a method you do not want to study deeply. Understand its question and contribution, then ask the model to find other treatments of the same or neighboring problems. Compare what each method reveals, its costs, and what it leaves unresolved. These connections can lead from a paper whose method does not appeal to you toward one using a method you prefer. Return to foundational reading when gaps prevent understanding it.

A preferred method supplies motivation to learn; its fit to a problem still depends on assumptions, evidence, and results. Retain important work that could change your judgment even when it does not match your interests. Models can retrieve, organize, explain, and compare. If you go on to make research claims, modify a method, or test results, deepen the learning into the relevant critical proofs, experiments, and counterexamples.

III. Everyday learning

If your purpose is everyday learning and broader understanding, reading an unfamiliar field's papers can usually focus on valuable, important abstractions and final conclusions. Ask the model what phenomena the paper represents through which objects and relationships, why that abstraction is useful, what conclusions it reaches, and how these connect to things you already know.

Also understand its key conditions, scope, and limitations. Use a concrete example and a case where it does not apply to check understanding. Technical details that do not serve the current purpose can be left aside, while keeping the authors' findings, the model's explanation, and your own associations distinguishable. This achieves understanding of important ideas and conclusions.

When an abstraction is especially fruitful, or you intend to make a practical decision from it, examine the corresponding argument, evidence, and foundations more closely. Everyday reading can already be a complete learning purpose.

Supplementary resources: disciplinary references and research methods

These libraries are supplementary resources for the teacher or model to read and select as needed. When useful, have the model download and connect them to the teaching and research environment:

The first enriches disciplinary understanding; the second supports concrete work. The teacher selects what the question needs.

Supplementary resources: scholarly writing and publication

For reading papers, developing arguments, or preparing submissions, ask the teacher or model to use these public methods:

The model selects methods for the actual writing question and connects the needed tools.