Join / Collaborate

Contact

We welcome postdoctoral researchers, PhD candidates, academic collaborators and industry partners interested in scientific learning and intelligent systems.

Xu Lab · How we research, learn, and grow

Get to know how we work together.

Before getting in touch, explore our approach to research and mentoring. Our charter sets out both the support members can expect and the responsibilities we share.

Explore our culture Read the full Lab Charter

Start a conversation

Find a way to work together.

A useful first conversation starts with the question you want to answer, the evidence available and the contribution you hope to make.

Scientific collaboration

Connect a question with a method.

For academic and industry teams exploring representation learning, model architectures, scientific data or the mathematical questions behind discovery.

What you can bring
Bring a well-defined question, your domain perspective and an idea of how progress could be assessed.
In your first email
  • The scientific question and where current methods fall short.
  • A short account of the evidence or theory available, with relevant public papers.
  • The expertise you would contribute and the part you would like to explore together.
A possible next step
If the interests align, an initial discussion can narrow the question and identify a first joint task.

Suggested email subjectXu Lab | Scientific collaboration

Email about collaboration Explore our research

Experimental partnership

Turn a testbed into informative evidence.

For experimental teams working with molecular interactions, biomaterials, delivery systems, proteins, peptides or other demanding scientific systems.

What you can bring
Bring knowledge of what can be measured or perturbed, the practical constraints and the alternatives an experiment could distinguish.
In your first email
  • The system, scientific question and decision the evidence should inform.
  • Available measurements, controls and feasible perturbations, including important constraints.
  • The experimental contribution you can make and the computational support you seek.
A possible next step
An exploratory discussion can define a useful measurement or pilot study and agree how its evidence would be assessed.

Suggested email subjectXu Lab | Experimental partnership

Discuss an experimental partnership Explore evidence engineering

Join the lab

Bring curiosity and a question to pursue.

For postdoctoral researchers, PhD candidates and other prospective students interested in model architectures, learning algorithms, evidence acquisition, or the design of scientific and intelligent systems.

A direction you can help shape
We welcome proposals whose central contribution is a new method or model architecture, developed through a scientific question. Tell us which direction you would like to shape and how it connects with our work.
What you can bring
Show how you approach a difficult question, learn unfamiliar methods and make your own contribution clear.
In your first email
  • A brief introduction and CV, including the role or programme you are considering and your intended start date.
  • A research question or direction that interests you and why it connects with the lab.
  • One work sample you can share, such as a report, paper, code example or project summary, with your contribution explained.
Ask about current opportunities
Please ask about openings, funding and application timing for your intended role or programme.
A possible next step
If there is a suitable opportunity, a discussion can explore the direction you want to develop, potential collaborators in the group, and the data, experimental or computational support a project would need.

Suggested email subjectXu Lab | Prospective researcher

Enquire about joining the lab Meet the team

Research asset use

Build on a specific piece of work.

For readers with a question about a published paper or a research resource linked from it, including how to reproduce or extend a method.

What you can bring
Bring a precise use case and enough context to identify the relevant paper, resource or technical question.
In your first email
  • The paper DOI or resource URL, and the version if one is provided.
  • What you want to understand, reproduce or extend, and what you have already tried.
  • A small example or exact error message when relevant; use material you are able to share.
A possible next step
A reply can clarify the relevant method or point to available material. Any access or reuse request depends on the specific resource and its terms.

Suggested email subjectXu Lab | Research asset enquiry

Ask about a paper or resource Browse publications

Primary contact

Shidang Xu

Address
No. 777, Xingye Avenue, Guangzhou, 510006
Book an appointment
Bird’s-eye view of the library.
Library · aerial view

Join / Collaborate

Bring a scientific question worth stress-testing.

A curved teal perforated panel, a navy rod framework and blue stacked forms are joined by amber connectors, leaving an open gap at the front.

A conceptual illustration of complementary expertise shaping a shared scientific question. Distinct components connect into an open structure with room for further contributions.

Our vision

Understanding the World. Pioneering the Future.

We seek to deepen our understanding of the world, recognise questions of growing importance, and cultivate researchers who can lead the future of their fields.

Scientific foresight

We look for early signals of change, connect developments across fields, see the structures behind phenomena, and question assumptions that limit our thinking. Our horizon spans decades and, where meaningful, centuries. What problems and constraints will become important? What knowledge could open new directions?

Evidence and intellectual leadership

We follow evidence in pursuit of reality and truth, remaining ready to revise our understanding. Thinking about the future means stating assumptions and turning them into questions we can investigate today. Intellectual leadership grows from worthwhile questions and rigorous understanding.

Cultivating domain intelligence system architects

One path is to cultivate domain intelligence system architects: researchers who connect model architecture, infrastructure, and the design of scientific and intelligent systems. Starting with worthwhile problems, they define objective functions (what to optimise), search spaces (what possibilities to explore), and evaluation criteria (how to judge progress).

For example, asking how much effective intelligence can be produced per unit of energy could open a new direction in energy-efficient intelligence. Our vision leaves room for important problems and research methods that have yet to emerge.