Lab Charter

This charter sets out our shared aims, educational commitments, and operating principles. It applies to the Principal Investigator (PI), Project Leaders (PLs), students, and other lab members. Each member has responsibilities appropriate to their role, and everyone helps maintain an environment that supports high-quality research and sustained growth.

1 | Purpose and Principles

Creating New Capabilities for Understanding and Discovery

We seek to expand humanity’s ability to understand complex systems, obtain reliable knowledge, and create new things. We investigate important scientific questions and explore more powerful ways of conducting research: how to represent problems, organize information, generate evidence, learn underlying patterns, and turn understanding into new designs and actions.

We are particularly interested in questions that remain insufficiently understood yet may have lasting significance. We hope to answer existing questions, discover new questions worth asking, identify critical gaps in current methods and knowledge, and establish research frameworks that can continually advance understanding.

By connecting artificial intelligence, mathematics, computation, and domain science, we aim to create concepts, methods, and systems of world-leading quality. We want to bring phenomena and objects that have been difficult to observe, explain, predict, or design within reach of systematic study, verification, and creation. Domain-specific intelligence, model architectures, infrastructure, and the design of scientific and intelligent systems are important paths toward this aim.

We value understanding problems from fundamental principles and organizing knowledge across levels into systems that work. We want our research to produce results and methods while also building capabilities that others can reuse, extend, and carry forward. A discovery should inspire new questions; a method should support multiple studies; a system should help more people make reliable judgments. Those who follow should be able to build on our work.

Through this process, members should grow into researchers who can ask important questions, understand fundamental principles, develop effective methods, assess research value, and independently advance their work. We hope to generate new knowledge and tools, and to develop people who can identify directions, build coherent research frameworks, and create possibilities for future research.

Our Principles

• Let important scientific questions guide method development. Explain the substantive difficulty a new method addresses, why it is needed, and what understanding or capability it could change.
• Actively discover and define problems. Look beyond existing questions and established evaluation criteria. Seek important unknowns, and examine whether objective functions, representations, search spaces, and evaluation criteria capture goals worth pursuing.
• Develop problems and approaches together. Encourage people with different backgrounds and expertise to participate in problem definition, theoretical analysis, system construction, and experimental validation, allowing the question and the approach to shape one another.
• Let methods and evidence constrain each other. Connect theory, computation, experiments, and feedback from real problems. Investigate why methods work and when they fail, and revise judgments promptly in light of evidence.
• Build capabilities that accumulate. Value the quality of concepts, code, data, tools, and infrastructure so that research outputs can be reused, combined, extended, and developed further, beyond a single result.
• Develop independent judgment through building together. Share design choices, reasoning, and lessons from failure. Respect different areas of expertise, and help members progressively learn to ask important questions, carry out independent work, and develop their own research directions.

2 | Student Development

Nurturing Curiosity, Learning to Think Independently, and Growing into Independent Research

We want graduate training to help students actively explore the world, understand problems deeply, and continue developing their abilities. Independent and critical thinking are essential foundations. Through guided learning, candid discussion, and practical research, the lab should help students formulate their own questions, examine existing understanding, and progressively learn to conduct research independently.

Begin with independent thinking and deep understanding. We encourage students to start from basic concepts and principles, understand a problem’s origins, assumptions, and logical relationships, explain what they learn in their own words, and apply that understanding in new situations. When presented with a conclusion, they should be able to ask: How was it reached? What assumptions does it depend on? What other explanations are possible? What evidence would undermine it? Guidance and feedback should help students identify gaps in understanding and pursue further questions, checks, and reasoning.

Develop critical thinking that is grounded and open to correction. Students should learn to distinguish facts, assumptions, inferences, and opinions; assess the reliability of sources and evidence; and examine whether conclusions exceed what the evidence supports. Answers from papers, experts, the PI, peers, and AI should all be judged on their supporting reasons and evidence. We equally value examining our own views: acknowledging what we do not know, accepting counterexamples, and revising our thinking when new evidence appears. The lab should take reasoned disagreement seriously and make questioning a normal part of learning.

Protect and rekindle curiosity. Students may join with interests that are not yet clear and questions that are still taking shape. Through stimulating reading, discussion, observation, and small exploratory studies, we should help them rediscover the pleasure of noticing a question, understanding something, and gaining a new insight. Mentors should pay attention to students’ spontaneous questions, make room for appropriate exploration, and help them develop questions they want to keep pursuing, starting from “What else do I want to understand?” Exploration need not immediately produce a paper, but it should have opportunities to lead to new understanding and further attempts.

Learn to choose information deliberately and manage attention. Amid abundant and continuous information, students should progressively learn to select inputs according to their own questions and learning goals, assess quality, relevance, and limitations, and preserve uninterrupted time for careful reading, reasoning, and practice. We encourage them to turn useful information into their own explanations, questions, connections, or tests, and to reflect regularly on which inputs deepen understanding and which simply consume attention. The lab should provide good materials and examples of learning, supporting independent and sustainable learning habits.

Make thinking itself part of shared learning. The PI, PLs, and students should share how questions emerge, approaches are selected, arguments develop, and evidence changes judgments. Discussion should reveal the key steps in reasoning and make room for failed attempts and unfinished ideas. Mentors should provide necessary knowledge and examples while using questions and feedback to preserve space for students to reason independently. Members deepen their understanding and learn different ways of thinking by explaining, comparing, and questioning ideas together.

Exercise active judgment in research and the use of AI. Real projects should help students connect problem definition, understanding of principles, method design, evaluation of evidence, and interpretation of results. AI can broaden thinking, support learning, and improve efficiency. Students should still be able to explain key choices, verify important supporting evidence, and take responsibility for the conclusions they adopt. We value the ability to ask worthwhile questions, connect different bodies of knowledge, test answers, and adjust actions in response to feedback. We also continually develop the disciplinary foundations, methods, and practical skills that support these judgments.

Build independence from different starting points. Students differ in their preparation, experience, interests, and goals. Training should provide tasks, challenges, and support that match current needs, with increasing autonomy as abilities develop. We look at actual changes in understanding, judgment, initiative, and work quality over time. We recognize sustained effort and accumulation, and allow for setbacks and revisiting earlier learning. Independent research can develop from understanding and completing a well-defined task toward designing approaches, assessing evidence, asking questions, and forming a research direction. Recognizing limitations, seeking help promptly, and collaborating effectively are also part of independence.

Research roles and career directions

Through real research, we help students develop the ability to frame important questions, design learning and exploration methods, gather and judge evidence, understand scientific mechanisms, and build systems and tools for sustained discovery.

Each student can develop a combination of strengths that reflects their interests, background and career goals, and adjust it as they gain experience. Independent thinking, curiosity, research integrity, collaboration and continued learning are foundations for every role.

Research roles we emphasize

• R01 · Scientific Problem Framer. Identify important unknowns and turn broad interests into worthwhile, testable research questions.
• R02 · Learning Objective Designer. Design learning signals, optimization objectives and constraints; decide what a system should pursue.
• R03 · Search Space Designer. Define the variables, candidates and combinations to explore, including their feasible ranges and boundaries.
• R04 · Scientific Representation Designer. Choose variables, coordinates, scales and representations that make key relationships easier to understand and learn.
• R05 · Model Architecture Designer. Design how information interacts, propagates, aggregates and updates within a model.
• R06 · Evidence and Experiment Designer. Decide which observation, simulation or experiment would distinguish explanations and reduce a key uncertainty.
• R07 · Scientific Evaluation Designer. Establish metrics, baselines, controls and tests to judge whether results represent real scientific progress.
• R10 · Scientific AI Systems Architect. Organize models, data, simulations, experiments and human judgment into a coordinated research process.
• R11 · AI Research Infrastructure Designer. Build the data, computing, training, tracking and reproducibility environment that supports sustained research.

Further research strengths to develop

• R08 · Mechanistic Researcher. Connect theory, computation and experiment to develop mechanistic explanations that can be tested.
• R09 · Mathematical and Computational Methods Researcher. Abstract structure from scientific problems and develop mathematical, inference, optimization and numerical methods.
• R12 · Research Software Developer. Turn scientific methods into correct, clear, maintainable software that others can reuse.

These roles describe areas of expertise that students can develop. Students can focus on one or two while understanding related parts of research; specific development goals should be agreed together, taking account of the current stage, interests and support needed.

From capabilities to careers: example pathways

Students can choose one or two capabilities as their main focus, develop expertise through concrete research contributions, and progressively gain experience relevant to their career goals.

These combinations can change with interests and experience. Students and mentors determine specific development goals together.

We hope students leave the lab with a lively curiosity about the world, habits of deep thinking, independent judgments that can withstand scrutiny, and the capacity to keep learning and creating. As knowledge, tools, and information environments change, these abilities should help them continue developing and choose for themselves the questions and paths worth pursuing.

3 | Shared Responsibilities and Decisions: Direction, Shared Thinking, and Accountability

We want a way of working together that enables us to discover questions, discuss them deeply, and act effectively. The PI, PLs, and students have different responsibilities and all contribute to scientific judgment. Leaders should create worthwhile research directions and conditions for growth; members should contribute their thinking and take responsibility for the work they undertake. Research quality, student development, and a healthy team depend on everyone taking appropriate responsibility for judgments, actions, time commitments, and accurate records.

The PI provides direction, intellectual guidance, and development opportunities. The PI’s central responsibility is to help the team identify and investigate more valuable questions, stay attentive to developments at the research frontier, judge which directions merit long-term investment, and identify the team’s distinctive contribution. The PI should lead research with the potential to make original contributions, open new directions, and shape future developments. They should preserve room for exploration while making necessary choices so that the team can build sustained expertise around important questions.

The PI should share their thinking: why a question matters, what is missing from current understanding, how to choose among approaches, and how new evidence changes a judgment. Sharing concepts, papers, reasoning, and research experience should help students develop deep understanding and independent thought. The PI should also actively create opportunities for valuable collaborations, exchange across disciplines, and scholarly discussion; allocate resources thoughtfully; develop PLs; and address organizational issues that affect research and student growth.

The PI assesses readiness and support needs before assigning responsibility. Assignments should reflect the knowledge and skills members demonstrate in practice, their follow-through on commitments, habits of proactive communication, manageable workload, and ability to coordinate and lead. For members whose work is less familiar, or for new responsibilities, clearly bounded tasks can provide a starting point for assessing readiness through practice and feedback.

Work assignments should specify expected outputs, decision authority, available resources, necessary guidance, and timing, so that responsibility, authority, and support are aligned. Tasks and autonomy should be adjusted as members grow or circumstances change. Assessment should support development and appropriate work arrangements, without turning performance at one stage into a fixed judgment of a person’s ability or responsibility.

PLs turn research direction into thoughtful progress. PLs should understand the scientific reasoning behind a project, its current evidence, intermediate goals, and main risks. They should help members connect specific tasks to the broader question and guide discussions toward explaining reasons, comparing approaches, and examining evidence. Project coordination and student development should reinforce each other, so that progress brings deeper understanding and stronger capabilities.

PLs should maintain attention to their projects, act on problems, follow through on commitments, and respond to difficulties. When work is blocked, they should investigate causes and coordinate support. When PI involvement is needed, they should explain the facts, options, recommendation, timing, and help required. PLs are responsible for accurate and timely reporting and should follow important matters through to resolution, an agreed change of plan, or a completed handover. Accountability includes acknowledging and correcting shortcomings in their own judgment or coordination.

Students take active responsibility for their work and learning. Students should understand the questions and tasks they undertake, bring their own thinking to discussions, explain their understanding, evidence, and uncertainties, and turn feedback into learning and action. In planning, they should be honest about existing work, limits of their abilities, and support needs. They should clarify unclear expectations and promptly explain their current understanding, attempts, and need for help when difficulties arise.

Students may differ in their preparation, pace of growth, and the scope and number of tasks they undertake. The lab maintains common expectations of honesty, care, responsibility, and timely communication. Students should remain responsible for work they have accepted, keep necessary records, proactively communicate results, and progressively take on more responsibility for research design and progress as their abilities develop.

Support collaboration through clear communication and reasonable commitments. Email is the lab’s primary work communication channel and takes priority over WeChat. The PI, PLs, and students should check email regularly on their agreed working days. Emails that clearly require a response or action from the recipient should receive a response, normally within two working days: either a substantive answer or an update explaining the current situation and when further feedback will be provided. General announcements, shared reading materials, and messages copied for information do not require individual acknowledgment.

WeChat supports convenient communication and reminders; it is not a formal channel on which a response is guaranteed. Matters requiring confirmation, decisions, task commitments, or a record of an important agreement should be raised or confirmed by email. For genuinely urgent matters, use a phone call or another mutually agreed direct channel to establish whether the person can respond in time.

Response periods are measured in the member’s agreed working days, excluding weekends, public holidays, and leave that has been communicated. Members may set aside specific times to handle messages and preserve uninterrupted time for experiments, reading, writing, and deep thinking. The time a message is sent does not create an expectation of an immediate response. Travel, extended experiments, or other arrangements that may affect availability should be communicated in advance where possible; unexpected circumstances can be explained as soon as conditions allow.

Make explicit commitments after understanding the task and workload. The person requesting work should explain its purpose, expected output, needed completion time, and relevant conditions. The person undertaking it should propose a feasible plan based on their abilities, existing work, and support needs. After acknowledging receipt, they may first clarify questions or assess the workload before agreeing on a completion date. If completion time cannot yet be estimated, they should specify when they will provide a further assessment.

Every important task should have a clear owner, completion criteria, and date. The PI’s reviews, feedback, and decisions should also be included in these arrangements. New tasks or changes in research expectations should be accompanied by a discussion of priorities and, where necessary, adjustments to existing deadlines. For uncertain research work, agree on assessable intermediate outputs and review points, allowing reasonable changes of approach in light of evidence.

Coordinate promptly when others’ progress is affected. If a reply or decision is blocking the next step, the requester should explain the specific impact and when a response is needed. If action is needed sooner than the normal response period, confirm availability through direct contact and adjust priorities or arrange other support where necessary. Urgency should be judged by actual consequences and time constraints.

Update plans proactively when delay becomes likely. Wherever possible, members should explain difficulties, consequences, attempts already made, and support needs before the agreed deadline, and propose a next step. Those involved should agree on revised arrangements and retain a necessary record of changes, so that people depending on the work are not left waiting. When work is completed, its results and any unresolved issues should be communicated proactively.

Address missed commitments through remedy and sustained improvement. An occasional omission or delay should first prompt a reminder, an explanation of the cause, and remedial action. If problems recur, those involved should examine whether tasks are clear, workloads reasonable, support sufficient, and communication and reminders effective.

If someone repeatedly misses deadlines without reasonable explanation, remains unresponsive for extended periods, or does not implement agreed improvements despite clear expectations, necessary support, and feedback, there should be a formal discussion establishing improvement expectations and a review date. Where necessary, task scope, project roles, or management responsibilities may be adjusted. Adjustments should be based on facts, take account of the person’s explanation, and allow reassessment following improvement.

These expectations apply equally to the PI and PLs. When delays arise from late guidance, review, or decisions, the responsible person should explain and address them, while adjusting affected members’ plans. Assessments of responsibility should consider actual work, follow-through, and how problems are handled.

Research integrity is fundamental to our work together. All members are responsible for the data generation, processing, interpretation, and reporting in which they participate. Fabricating or falsifying data, plagiarism, concealing key results that affect conclusions, and misleading selective presentation are prohibited. No one may demand untruthful research results to meet a deadline, support a submission, or satisfy an expected outcome.

Experimental research should retain raw data, sample and experimental conditions, instrument outputs, and necessary procedural records. Computational research should record data sources, code, parameters, and processing steps. Data exclusions, image processing, and other operations affecting interpretation should retain their rationale and records. Failed experiments, negative results, anomalies, and errors should be recorded and discussed honestly. Simulated or generated data should be clearly identified as such.

The PI and PLs should provide necessary guidance on research integrity and data management and arrange checks appropriate to the research risks. Questions about data or conclusions should be clarified promptly, preserving original records and suspending use of the results where necessary. Suspected misconduct should be investigated and handled under applicable institutional procedures, allowing those involved to explain, distinguishing honest error from deliberate fabrication, and protecting members who raise questions or report concerns in good faith.

Keep discussion open and make authority and decisions explicit. Scientific judgments should be tested by evidence and argument. Members may question anyone’s views, including the PI’s. At project initiation, clarify who leads the work, the scope of their authority, and matters requiring joint discussion. Important changes to research direction, core scientific claims, major resource commitments, external commitments, or submission plans should be decided by the PI or an explicitly authorized lead.

When action is needed, the responsible lead should make a decision after adequately considering relevant views, explain the reasoning, and specify responsibilities, timing, and support for the next steps. The team should continue attending to new evidence after implementation. Changes in direction or conditions should prompt timely discussion and corresponding adjustments to tasks and plans.

4 | Daily Work and Collaboration: Initiative, Deep Engagement, and Reliability

We want members to develop ways of working that support sustained research: knowing which questions matter now, managing time and attention, moving work forward proactively, and collaborating reliably. The lab should provide consistent guidance, resources, and conditions for working together so that these abilities develop through practice.

Set priorities around the research question. Members should understand how current work relates to the broader research goal and translate larger goals into actionable reading, derivation, experiments, computation, or writing. Task planning should account for learning needs, dependencies, resources, and existing workloads. The PI and PLs should help members make choices and avoid taking on so many tasks that none can be pursued deeply.

Establish daily arrangements that enable collaboration. Subject to applicable university requirements, regular work together on weekdays is the norm, with reasonable individual arrangements. At this stage, shared presence and availability should be coordinated by project, with engagement, progress, and reliable collaboration as the main criteria. There is no lab-wide clock-in requirement or routine requirement for evening or weekend overtime. Absences, remote work, and schedule changes should address affected collaboration and handovers. Leave follows applicable policies; unexpected circumstances should be communicated as soon as conditions allow.

Use support and adjustments with a defined review period for repeated delays. For members with repeated significant delays or persistent difficulties at important milestones such as proposal or midterm reviews, the PI and PL should examine concrete records to understand the causes. They should consider task difficulty, workload, guidance and resources, communication, and daily arrangements, and establish near-term goals, necessary support, and review dates. Problems at important milestones should receive direct help addressing the scientific question, study design, evidence, and presentation.

Where needed, members may be asked to make a brief daily update on agreed working days for a defined period. The update should cover progress, supporting evidence, difficulties and help needed, and the next working day’s priorities. Records should use existing project pages wherever possible. The PL should read them and respond promptly to matters requiring attention; important questions of direction, decisions, and training should be brought to the PI. When there are no new results, members should accurately describe their attempts and what they learned.

If discussion establishes that an unstable routine or insufficient shared working time is an important cause, the PI may specify temporary arrival times and shared working periods after hearing the member’s views, taking account of classes, experiments, and reasonable individual needs. Increased reporting frequency and fixed attendance times should each be justified by actual needs and accompanied by appropriate support.

At the start of an adjustment, specify review dates and the conditions for returning to normal weekly updates and greater scheduling autonomy. Reviews should assess improvements in progress, communication, and working habits. Normal arrangements should be progressively restored as things improve; continuing difficulties should prompt a fresh examination of causes and support. Any extension should have a stated basis and a new review date. Delay does not automatically lead to longer daily hours or evening or weekend overtime. Effects caused by late guidance, reviews, decisions, or resources should be addressed by the responsible person, with corresponding changes to affected members’ plans.

Protect time for deep thinking. Reading, reasoning, design, experiments, and writing require sustained attention. Members should actively reduce aimless information consumption and frequent switching between tasks. The PI and PLs should also plan meetings and unanticipated requests thoughtfully to support uninterrupted work. Reviews should examine changes in understanding, evidence, and work quality.

Preserve room for exploration. Members may propose new questions and small exploratory studies, discussing their significance, required effort, and relation to current research with mentors. The lab explicitly accepts that some worthwhile learning and exploration will not produce papers in the near term, and makes room for this work. Exploratory tasks should have a clear question, a reasonable scope of effort, and review points. When positive results have not emerged, members should explain what they learned, which possibilities they ruled out, and how they will judge the next step. Important new leads may justify a proposal to expand exploration or adjust the main project.

Use resources responsibly and share common duties. Use of shared equipment, computing resources, funds, and others’ support should be coordinated in advance, with status updated afterward. Shared maintenance, onboarding, and mutual assistance should have clear arrangements, with periodic checks that burdens are reasonably balanced. The PI and PLs should account for these contributions in assessing workload and contribution.

Maintain sustainable effort. The lab values serious work and supports necessary rest and recovery. Concentrated work ahead of important deadlines should be coordinated in advance. Persistent overload, prolonged difficulties, or repeated last-minute rushes should prompt a review of tasks, methods, and support. Members may raise difficulties promptly and seek adjustments.

5 | Progress Reporting and Scholarly Exchange: Feedback for Thinking, Support for Progress

Reporting and discussion should help the team understand the state of research, improve judgments, and resolve obstacles. Explaining the research process is also part of learning to think. The lab should give communication a clear purpose and keep preparation proportionate to its value.

Use brief weekly records to show what has actually changed. Students should update the agreed project page on the lab’s Google site once a week, using that record directly in discussions with their PL. Updates should explain the current question, new work and understanding since the previous update, supporting materials, remaining difficulties, next steps, and support needed. If there are no new results, state the reasons, attempts made, and the judgment needed next. Existing records may be cited for background; past results must not be presented as newly achieved progress. Necessary reminders, requests, and formal confirmations should still be sent by email under Section 3.

Match routine guidance to members’ needs. PLs should use weekly records to support progress through timely short discussions or written feedback. Each student should normally meet directly with the PI every 2–4 weeks to discuss research, development, and difficulties. Contact may be more frequent during onboarding, topic selection, repeated difficulties, or major transitions. The PI and student should both maintain these commitments; changes due to travel or other reasons should include rescheduled contact or alternative support. PL guidance and direct PI involvement together support student development. Important obstacles should prompt communication without waiting for the next regular meeting.

Organize whole-lab scholarly discussion around worthwhile questions. The PI, PLs, and students may all propose topics, including research judgments, papers, lessons from failure, methods and technical tools, and presentation rehearsals. Each session should have a clear topic, a main presenter, and a facilitator, with enough background for members of different experience levels to understand and participate.

Presenters should explain their reasoning, evidence, limitations, and the feedback they seek. Members may ask clarifying questions, propose alternative explanations, and suggest tests. There is no requirement for everyone to speak in turn at every session, and participation is not judged by the number of comments made in one meeting. The PI and PLs should help members learn how to form and revise judgments. Members can still organize valuable discussions when the PI is unavailable. The format, duration, frequency, and review arrangements are set out in the companion Operating Notes.

Use PL Monthly Reports to support team decisions. Each PL submits a monthly report to the PI on the projects or team they lead and takes part in a PI–PL discussion. Reports should present new work and understanding during the month, key evidence, main risks, members’ support needs, and matters requiring PI decisions. PI–PL discussions should focus on priorities across projects, resources, and collaboration, and specify who will follow up, when, and with what support. Monthly materials should draw on existing project records wherever possible. Submission arrangements and formats are specified in the Operating Notes.

Reduce duplicate preparation and retain useful conclusions. The same research records can support routine guidance, progress reports, and manuscript preparation. Routine exchanges should use a concise format that makes the issue clear, while formal presentations provide opportunities for more extensive training in structure and delivery. Meetings should conclude by clarifying the main conclusions, disagreements, and next steps, with shared records updated promptly. Participants should be informed in advance of recordings or AI-generated minutes, and access should reflect the sensitivity of the content.

6 | Research Quality and Publication: Valuable Contributions Built on Reliable Evidence

We want research to make credible, clear contributions to important questions that can support further work. Quality should inform topic selection, design, execution, analysis, writing, and public communication, alongside the research integrity requirements in Section 3.

Clarify research value and key uncertainties before committing resources. Projects should explain where the question comes from, what is already understood, the expected contribution, and the assumptions most in need of testing. Depending on the project, prioritize small analyses or experiments that test the central idea, with appropriate conditions for continuing, adjusting, or stopping. Project reviews should respect evidence and consider research value, learning, effort invested, and remaining opportunities together.

Keep arguments proportionate to evidence. Members should examine whether comparisons are fair, controls adequate, conclusions supported, and results robust to reasonable alternative explanations. Computational, experimental, and theoretical work should use checks suited to their nature. Important conclusions should receive appropriate review by peers. Anomalies, failures, and negative results should inform research judgments.

Organize research early through writing and figures. As a project develops, members should progressively articulate the scientific question, main contribution, core figures, and gaps in evidence. Authors, the PL, and the PI should discuss key scientific judgments and the manuscript’s structure before developing a full draft. Figures should clearly identify the data, conditions, and basis of comparison. Conceptual illustrations should accurately distinguish established understanding from ideas still to be tested.

Make workable arrangements for review and submission. Materials submitted for review should state the feedback sought, and reviewers should confirm their arrangements for responding. Revisions should distinguish work necessary to support conclusions, additional work that would substantially strengthen the study, and questions that can be left for the future, with scope and timing adjusted accordingly. Submission strategy should consider quality, audience, research timeliness, and members’ development needs. Persistent obstacles to the original plan should prompt an explicit discussion of alternatives.

Discuss contributions, authorship, and collaboration early. Project participants should understand their expected contributions and revisit the discussion as the work changes. Authorship should reflect actual scholarly contributions, corresponding responsibilities, and applicable disciplinary and publication requirements. All authors should have an opportunity to review the manuscript and agree to the final content submitted. Members may raise questions about authorship promptly and should receive an explanation and appropriate help resolving them. Shared maintenance, tool sharing, and mentoring new members should receive appropriate recognition and rewards; whether they qualify as authorship contributions to a particular paper depends on their actual scholarly contribution to that paper.

Make research outputs understandable and usable. The lab supports, in principle, public preprints and reusable code, and encourages dissemination of methods and data that can be shared. Where permissions, coauthor agreements, and applicable agreements allow, the work required for release and reuse should be included in project plans. Relevant permissions and agreements should be confirmed before release, and necessary instructions, version information, and limitations should be provided. For restricted materials, explain the restrictions and feasible ways to verify or access the work wherever possible.

Create collaboration opportunities and coordinate them responsibly. Members may initiate scholarly exchanges and propose collaborations, keeping relevant colleagues informed of opportunities. Additional workloads, funding, equipment, data access, or external commitments should be discussed in advance with the person authorized to decide, clarifying goals, contributions, support, and timing.

Retain human judgment and responsibility when using AI. AI can assist learning, programming, figures, writing, and organizing materials. Members should understand key methods, verify references, code behavior, and important conclusions, and take responsibility for what they submit. Use of external tools should respect permissions governing the materials involved. Disclosure of AI-generated content should follow applicable requirements.

7 | Evaluation, Rewards, and Opportunities: Recognizing Contributions and Supporting Growth

Evaluation should help members understand their progress, contributions, and next goals, while helping the team allocate resources and opportunities fairly. It should be grounded in actual work and consider research quality, individual growth, and the team’s long-term development.

Understand contributions broadly. Evaluation should consider the value of questions and methods, reliability of results, research progress and follow-through, development of independent judgment, and collaboration and shared work. Papers, theoretical analyses, experimental platforms, code, data, tools, mentoring, and knowledge sharing can all make important contributions. Assessment should account for research difficulty, stage, and actual responsibilities, so that immediately visible outputs do not obscure long-term accumulation.

Hold useful, regular development reviews. The lab conducts a comprehensive evaluation every six months, drawing on concrete work over time to discuss research and team contributions. Individual growth and career discussions follow a separate schedule of once per half-semester. Members may give feedback on guidance and arrangements. Reviews should identify a small number of clear areas for improvement, with subsequent discussions checking whether support has been provided and actions taken.

Recognize active learning and positive influence on others. Proactive inquiry, willingness to strengthen knowledge and skills, careful follow-through on feedback, persistence in finding ways through difficulties, and helping peers sustain their curiosity and motivation should all be considered in rewards. Such contributions can be recognized before papers or other outputs emerge. Assessment should draw on concrete learning actions, improvement over time, and examples of helping others. It should respect different personalities and ways of communicating, and avoid judging solely by hours present, visible activity, or deference to mentors.

Reward useful sharing of techniques and tools. Sharing research methods, programming techniques, experimental experience, and tool use—particularly effective practices for AI-assisted learning and research—should count toward team contributions and reward decisions. Wherever possible, sharing should explain suitable problems, how the approach is used, its effects, limitations, and ways to check it, leaving materials that peers can try or reuse. Identifying errors, risks, or unsuitable uses is also valuable. Recognition should reflect practical benefit, learning value, and maintenance effort, rather than simply the number of sessions or a tool’s popularity.

Keep reward rules clear and explainable. The lab provides appropriate rewards and development support in light of actual contributions and available resources. Infrastructure, onboarding, shared resource maintenance, active learning, and sharing technical tools should all be considered. Eligibility, evaluation periods, criteria, and decision procedures should be explained in advance.

PLs provide concrete records of project and team contributions. Members may add examples of their work, growth, and support for others. The PI makes an overall assessment and explains the basis to the individual. Members may provide further information and request reconsideration of omissions, factual errors, or disputed assessments. We hope to turn the additional resources made possible by the lab’s development into better training conditions and member incentives. Specific reward arrangements depend on project funding conditions and applicable requirements.

Match opportunities with development support. Project leadership, external collaborations, conferences, and other opportunities should be arranged with attention to relevant abilities, interests, readiness, previous opportunities, and actual needs. Members should understand how they can develop toward greater responsibility. Everyone should receive guidance and learning opportunities appropriate to their training goals, with tasks and support adjusted as they grow.

Support different career paths. Helping students move toward a suitable next stage is an explicit educational responsibility of the lab. We value academic research and respect careers in industry, education, and other fields. The PI and each student should hold a growth and career discussion at least once per half-semester—at least twice each semester—which may be incorporated into an existing individual meeting. Discussions should cover interests and goals, changes in capabilities, support needed, and next steps, allowing choices to evolve with experience. Internships, exchanges, and external collaborations should be coordinated in advance in light of training goals, project plans, and applicable requirements, with clear responsibilities and handovers.

Concrete career directions

The capabilities described in Section 2 can support a range of future careers. Job titles and responsibilities vary across institutions, and some directions overlap. Students can choose, combine and revise their goals over time.

Career directions we explicitly encourage

• C01 · Professor or Faculty Member. Conduct research, teach and mentor students, and develop an academic direction.
• C02 · Principal Investigator. Set a research agenda, organize a team and collaborations, and take responsibility for research and training.
• C03 · Research Scientist. Pursue sustained research on scientific questions, methods and discoveries.
• C04 · Machine Learning Research Scientist. Investigate learning methods, scientific representations, model architectures and generalization.
• C05 · Applied Scientist. Translate research methods into verifiable improvements in industrial R&D or products.
• C13 · Machine Learning Engineer. Implement training, inference and model integration, and verify operational performance.
• C18 · AI Systems Architect. Coordinate system components, interfaces, feedback, performance and reliability.
• C19 · Laboratory Automation Engineer. Connect instruments, software, robotics and experimental workflows.
• C22 · R&D Lead. Organize technical direction, team collaboration and R&D outcomes.
• C24 · Scientific Entrepreneur or Technical Cofounder. Turn technical discoveries into products or services and build a team for sustained development.

Further career directions to explore

Scientific and domain research

• C07 · Computational Materials Scientist. Study relationships among material structures, properties, conditions and responses.
• C08 · Computational Biologist. Model biological systems and perturbations to develop testable biological explanations.
• C09 · Bioinformatics Scientist. Analyze sequence, omics and other data to answer specific biological questions.
• C10 · Molecular Design Scientist. Design and validate candidate molecules with desired properties or functions.
• C12 · Experimental R&D Scientist. Design and conduct experiments, connecting measurements, mechanistic judgment and R&D decisions.

Scientific computing, software and engineering

• C11 · Scientific Computing Researcher. Develop numerical methods, simulations and approaches to inverse problems.
• C14 · Research Engineer. Build research prototypes that turn complex ideas into testable experiments.
• C15 · Research Software Engineer. Develop, validate, maintain and extend research software.
• C16 · Scientific Data Engineer. Build processes for data collection, processing, quality checks and traceability.
• C17 · ML Platform Engineer. Build shared platforms for training, computing, experiment tracking and reproducibility.

Evaluation and product development

• C20 · AI Evaluation Researcher. Design evaluation tasks and tests, and study model capabilities and failure boundaries.
• C21 · Scientific AI Product Manager. Translate researchers’ needs into product goals, features and acceptance criteria.

We support members in exploring these paths and developing the specialist skills they need through courses, co-mentoring, internships and further training. Experimental R&D, laboratory automation, product development and entrepreneurship also require relevant practical experience and working conditions.

8 | Our Working Environment: Respect, Candor, and Mutual Support

High research standards and mutual respect should reinforce each other in the lab. Every member should be able to focus, receive reasonable support, ask questions, and continue learning through disagreement.

Treat every member with care and respect. Discussion should address specific questions, behavior, and evidence, with feedback that people can understand and act on. Humiliation, discrimination, harassment, and threats must not be used in management or communication. Members may differ in background, personality, and expression; the team should provide clear and fair opportunities to participate.

Respond to requests for help and differing views. Members may acknowledge gaps in understanding, report errors, and express disagreement. The PI and PLs should actively learn about members’ needs, paying particular attention to students who are not yet comfortable seeking help. Reasonable questions, honest reports of unfavorable results, and good-faith concerns about management should not lead to retaliation or unfair adverse treatment.

Provide accessible ways to address problems. Routine disagreements may be discussed among those involved or with help from a PL or the PI. Members may also directly use the university’s existing advice, support, or formal procedures. Matters involving the PI, conflicts of interest, or other circumstances unsuitable for internal handling do not require prior reporting to the PI before external help can be sought. Information should be shared only as needed, with specific channels provided in the Operating Notes.

Maintain safety and health together. Before undertaking experimental, equipment-related, data-related, or other relevant work, members should complete applicable training and understand operating conditions, risks, and how to seek help. Unfamiliar or unsafe procedures should prompt a pause and a request for guidance. Accidents, anomalies, and potential risks should be reported promptly; the responsible person should arrange a response and improve conditions.

Respect personal boundaries. Shared activities should consider members’ time, financial burden, and individual circumstances. The lab supports reasonable rest, and participation in social activities is voluntary. Health and personal difficulties may be discussed through appropriate channels, with work coordination focused on necessary support and arrangements.

9 | Shared Records and Continuity: Preserving Knowledge for Those Who Follow

The lab’s accumulated work should be understandable and usable by current members, collaborators, and those who join later. Good records and handovers support reliable research and reduce repeated searching, rework, and explanation.

Provide a clear shared entry point for materials. For each project, members should be able to find the current question, key decisions, research records, locations of data and code, progress materials, and the person responsible. The lab should agree on primary storage and version management practices, avoiding long-term reliance on personal devices or private chats for important information. Routine management tools should be straightforward to maintain and use.

Integrate record keeping into research. Preserve traceable connections among original materials, processing methods, versions, and conclusions. Progress reports and manuscripts should draw on shared sources wherever possible and state dates and status clearly. Backups, access permissions, and maintenance should have explicit arrangements, while protecting personal information and restricted research materials.

Help new members get started. Onboarding should introduce the lab’s principles, communication practices, training arrangements, basic tools, and ways to seek help, with necessary instruction and examples. When experienced members help mentor newcomers, clarify their expected effort and available support, helping new members progressively learn to use resources and advance work independently.

Plan graduation, departure, and handovers in advance. At an appropriate stage, members and the PI should discuss completion timing, the scope of work to finish, manuscript and material handovers, future plans, and sustainable ways to collaborate, based on university degree requirements and agreed training goals. Pursuit of higher-profile publication should have a defined scope of effort and review arrangements, so that continually raising journal targets does not leave graduation and project completion open-ended. Handovers should state the current status, material locations, known problems, and who will take over. Further work after graduation or departure should be separately agreed in scope, time, and support, with prior contributions continuing to be respected. The lab should help students complete necessary handovers and move toward a suitable next stage.

Improve shared resources over time. Valuable methods, tutorials, code, and lessons from failure can be organized into reusable team resources. Maintenance should receive recognition, and resources should be updated in response to use. We encourage members to carry the approaches to inquiry and collaboration they have developed into new environments.

10 | Using and Updating the Charter: Putting Principles into Practice

This charter explains our shared aims, educational commitments, and operating principles. New members should have an opportunity to read it, discuss it, and seek clarification. The PI and PLs should follow these commitments in everyday judgments and arrangements.

Support implementation with clear operating details. Reporting frequency, meeting arrangements, record templates, resource use, reward procedures, and related details may be specified in the companion Operating Notes. The charter and operating details should be consistent with applicable university and school policies. Changes should explain their reasons, scope, and effective date.

Review how arrangements work in practice. The lab should periodically review research progress, training support, communication burden, the collaborative environment, and allocation of opportunities. Members may suggest improvements. New rules should address clear needs; procedures that persistently fail to support research or development should be simplified, changed, or removed.

Revise transparently. The PI should organize discussion and make clear decisions on important revisions, retaining versions and a record of major changes. Changes affecting important member arrangements should be communicated in advance, addressing existing commitments and transition needs. The charter’s value lies in members being able to understand and practice it, and to help improve it through experience.