- Schools and departments
Physics and Astronomy
- Postgraduate study
Guidelines for writing a research proposal
- Undergraduate study
- Postgraduate prospectus
- Preparatory study for MSc courses
- PhD Opportunities
- Prizes for postgraduate Physics and Astronomy students
- Why physics at Sussex?
- People and contacts
The research proposal is central to your application to undertake doctoral study in the Department of Physics and Astronomy. You should read the following guidelines to ensure that your proposal includes the information we need to assess your application.
To support your application for a PhD place in the School of Mathematical & Physical Sciences, you should either:
- write a new project proposal in a specialised subject, which will appeal to our faculty, naming one or more preferred supervisors.
- write a general personal statement describing a broad topic of interest to you and our faculty. You should show how your areas of academic strength would benefit the topic. If you have more than one such topic, clearly address each one separately. You may indicate preferred supervisors and research groups.
- write a specific personal statement about why you are the right person for one of our advertised research projects. (Not all research groups advertise projects.)
As appropriate, you can
- explain your interest and motivation to carry out the research
- specify the questions you wish to investigate, including references to research literature
- indicate which methods and techniques are adequate to achieve the research aims, and state whether you are able to apply them or wish to develop skills in them.
Your document can be from half to four pages in length, as necessary. Please upload it to your application in pdf format.
Details of our Research Groups can be found on the research webpage .
In the Financial Information section of the online form, you should:
- Tell us which studentship you would like to be considered for. They are advertised on the scholarships web pages . Check your eligibility in the advertisement.
- Tell us if you have another way of funding your studies should we be unable to offer you a studentship.
- Tell us the name of your sponsor, or intended sponsor, if you will be funded by a third party. Inform us of any important deadlines.
Research Proposal Example/Sample
Detailed Walkthrough + Free Proposal Template
If you’re getting started crafting your research proposal and are looking for a few examples of research proposals , you’ve come to the right place.
In this video, we walk you through two successful (approved) research proposals , one for a Master’s-level project, and one for a PhD-level dissertation. We also start off by unpacking our free research proposal template and discussing the four core sections of a research proposal, so that you have a clear understanding of the basics before diving into the actual proposals.
- Research proposal example/sample – Master’s-level (PDF/Word)
- Research proposal example/sample – PhD-level (PDF/Word)
- Proposal template (Fully editable)
If you’re working on a research proposal for a dissertation or thesis, you may also find the following useful:
- Research Proposal Bootcamp : Learn how to write a research proposal as efficiently and effectively as possible
- 1:1 Proposal Coaching : Get hands-on help with your research proposal
FAQ: Research Proposal Example
Research proposal example: frequently asked questions, are the sample proposals real.
Yes. The proposals are real and were approved by the respective universities.
Can I copy one of these proposals for my own research?
As we discuss in the video, every research proposal will be slightly different, depending on the university’s unique requirements, as well as the nature of the research itself. Therefore, you’ll need to tailor your research proposal to suit your specific context.
You can learn more about the basics of writing a research proposal here .
How do I get the research proposal template?
You can access our free proposal template here .
Is the proposal template really free?
Yes. There is no cost for the proposal template and you are free to use it as a foundation for your research proposal.
Where can I learn more about proposal writing?
For self-directed learners, our Research Proposal Bootcamp is a great starting point.
For students that want hands-on guidance, our private coaching service is recommended.
Psst… there’s more!
This post is an extract from our bestselling Udemy Course, Research Proposal Bootcamp . If you want to work smart, you don't want to miss this .
You Might Also Like:
I am at the stage of writing my thesis proposal for a PhD in Management at Altantic International University. I checked on the coaching services, but it indicates that it’s not available in my area. I am in South Sudan. My proposed topic is: “Leadership Behavior in Local Government Governance Ecosystem and Service Delivery Effectiveness in Post Conflict Districts of Northern Uganda”. I will appreciate your guidance and support
GRADCOCH is very grateful motivated and helpful for all students etc. it is very accorporated and provide easy access way strongly agree from GRADCOCH.
I am at the stage of writing my thesis proposal for a masters in Analysis of w heat commercialisation by small holders householdrs at Hawassa International University. I will appreciate your guidance and support
please provide a attractive proposal about foreign universities .It would be your highness.
comparative constitutional law
Kindly guide me through writing a good proposal on the thesis topic; Impact of Artificial Intelligence on Financial Inclusion in Nigeria. Thank you
Submit a Comment Cancel reply
Your email address will not be published. Required fields are marked *
Save my name, email, and website in this browser for the next time I comment.
- Print Friendly
- Free Samples
- Premium Essays
- Editing Services Editing Proofreading Rewriting
- Extra Tools Essay Topic Generator Thesis Generator Citation Generator GPA Calculator Study Guides Donate Paper
- Essay Writing Help
- About Us About Us Testimonials FAQ
- Physics Research Proposal
- Samples List
An research proposal examples on physics is a prosaic composition of a small volume and free composition, expressing individual impressions and thoughts on a specific occasion or issue and obviously not claiming a definitive or exhaustive interpretation of the subject.
Some signs of physics research proposal:
- the presence of a specific topic or question. A work devoted to the analysis of a wide range of problems in biology, by definition, cannot be performed in the genre of physics research proposal topic.
- The research proposal expresses individual impressions and thoughts on a specific occasion or issue, in this case, on physics and does not knowingly pretend to a definitive or exhaustive interpretation of the subject.
- As a rule, an essay suggests a new, subjectively colored word about something, such a work may have a philosophical, historical, biographical, journalistic, literary, critical, popular scientific or purely fiction character.
- in the content of an research proposal samples on physics , first of all, the author’s personality is assessed - his worldview, thoughts and feelings.
The goal of an research proposal in physics is to develop such skills as independent creative thinking and writing out your own thoughts.
Writing an research proposal is extremely useful, because it allows the author to learn to clearly and correctly formulate thoughts, structure information, use basic concepts, highlight causal relationships, illustrate experience with relevant examples, and substantiate his conclusions.
- Research Proposal
Examples List on Physics Research Proposal
- TERMS & CONDITIONS
- COOKIES POLICY
Have a language expert improve your writing
Run a free plagiarism check in 10 minutes, generate accurate citations for free.
- Knowledge Base
- Starting the research process
- How to Write a Research Proposal | Examples & Templates
How to Write a Research Proposal | Examples & Templates
Published on October 12, 2022 by Shona McCombes and Tegan George. Revised on November 21, 2023.
A research proposal describes what you will investigate, why it’s important, and how you will conduct your research.
The format of a research proposal varies between fields, but most proposals will contain at least these elements:
- Research design
While the sections may vary, the overall objective is always the same. A research proposal serves as a blueprint and guide for your research plan, helping you get organized and feel confident in the path forward you choose to take.
Table of contents
Research proposal purpose, research proposal examples, research design and methods, contribution to knowledge, research schedule, other interesting articles, frequently asked questions about research proposals.
Academics often have to write research proposals to get funding for their projects. As a student, you might have to write a research proposal as part of a grad school application , or prior to starting your thesis or dissertation .
In addition to helping you figure out what your research can look like, a proposal can also serve to demonstrate why your project is worth pursuing to a funder, educational institution, or supervisor.
Research proposal length
The length of a research proposal can vary quite a bit. A bachelor’s or master’s thesis proposal can be just a few pages, while proposals for PhD dissertations or research funding are usually much longer and more detailed. Your supervisor can help you determine the best length for your work.
One trick to get started is to think of your proposal’s structure as a shorter version of your thesis or dissertation , only without the results , conclusion and discussion sections.
Download our research proposal template
Prevent plagiarism. Run a free check.
Writing a research proposal can be quite challenging, but a good starting point could be to look at some examples. We’ve included a few for you below.
- Example research proposal #1: “A Conceptual Framework for Scheduling Constraint Management”
- Example research proposal #2: “Medical Students as Mediators of Change in Tobacco Use”
Like your dissertation or thesis, the proposal will usually have a title page that includes:
- The proposed title of your project
- Your supervisor’s name
- Your institution and department
The first part of your proposal is the initial pitch for your project. Make sure it succinctly explains what you want to do and why.
Your introduction should:
- Introduce your topic
- Give necessary background and context
- Outline your problem statement and research questions
To guide your introduction , include information about:
- Who could have an interest in the topic (e.g., scientists, policymakers)
- How much is already known about the topic
- What is missing from this current knowledge
- What new insights your research will contribute
- Why you believe this research is worth doing
As you get started, it’s important to demonstrate that you’re familiar with the most important research on your topic. A strong literature review shows your reader that your project has a solid foundation in existing knowledge or theory. It also shows that you’re not simply repeating what other people have already done or said, but rather using existing research as a jumping-off point for your own.
In this section, share exactly how your project will contribute to ongoing conversations in the field by:
- Comparing and contrasting the main theories, methods, and debates
- Examining the strengths and weaknesses of different approaches
- Explaining how will you build on, challenge, or synthesize prior scholarship
Following the literature review, restate your main objectives . This brings the focus back to your own project. Next, your research design or methodology section will describe your overall approach, and the practical steps you will take to answer your research questions.
To finish your proposal on a strong note, explore the potential implications of your research for your field. Emphasize again what you aim to contribute and why it matters.
For example, your results might have implications for:
- Improving best practices
- Informing policymaking decisions
- Strengthening a theory or model
- Challenging popular or scientific beliefs
- Creating a basis for future research
Last but not least, your research proposal must include correct citations for every source you have used, compiled in a reference list . To create citations quickly and easily, you can use our free APA citation generator .
Some institutions or funders require a detailed timeline of the project, asking you to forecast what you will do at each stage and how long it may take. While not always required, be sure to check the requirements of your project.
Here’s an example schedule to help you get started. You can also download a template at the button below.
Download our research schedule template
If you are applying for research funding, chances are you will have to include a detailed budget. This shows your estimates of how much each part of your project will cost.
Make sure to check what type of costs the funding body will agree to cover. For each item, include:
- Cost : exactly how much money do you need?
- Justification : why is this cost necessary to complete the research?
- Source : how did you calculate the amount?
To determine your budget, think about:
- Travel costs : do you need to go somewhere to collect your data? How will you get there, and how much time will you need? What will you do there (e.g., interviews, archival research)?
- Materials : do you need access to any tools or technologies?
- Help : do you need to hire any research assistants for the project? What will they do, and how much will you pay them?
If you want to know more about the research process , methodology , research bias , or statistics , make sure to check out some of our other articles with explanations and examples.
- Sampling methods
- Simple random sampling
- Stratified sampling
- Cluster sampling
- Likert scales
- Null hypothesis
- Statistical power
- Probability distribution
- Effect size
- Poisson distribution
- Optimism bias
- Cognitive bias
- Implicit bias
- Hawthorne effect
- Anchoring bias
- Explicit bias
Once you’ve decided on your research objectives , you need to explain them in your paper, at the end of your problem statement .
Keep your research objectives clear and concise, and use appropriate verbs to accurately convey the work that you will carry out for each one.
I will compare …
A research aim is a broad statement indicating the general purpose of your research project. It should appear in your introduction at the end of your problem statement , before your research objectives.
Research objectives are more specific than your research aim. They indicate the specific ways you’ll address the overarching aim.
A PhD, which is short for philosophiae doctor (doctor of philosophy in Latin), is the highest university degree that can be obtained. In a PhD, students spend 3–5 years writing a dissertation , which aims to make a significant, original contribution to current knowledge.
A PhD is intended to prepare students for a career as a researcher, whether that be in academia, the public sector, or the private sector.
A master’s is a 1- or 2-year graduate degree that can prepare you for a variety of careers.
All master’s involve graduate-level coursework. Some are research-intensive and intend to prepare students for further study in a PhD; these usually require their students to write a master’s thesis . Others focus on professional training for a specific career.
Critical thinking refers to the ability to evaluate information and to be aware of biases or assumptions, including your own.
Like information literacy , it involves evaluating arguments, identifying and solving problems in an objective and systematic way, and clearly communicating your ideas.
The best way to remember the difference between a research plan and a research proposal is that they have fundamentally different audiences. A research plan helps you, the researcher, organize your thoughts. On the other hand, a dissertation proposal or research proposal aims to convince others (e.g., a supervisor, a funding body, or a dissertation committee) that your research topic is relevant and worthy of being conducted.
Cite this Scribbr article
If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.
McCombes, S. & George, T. (2023, November 21). How to Write a Research Proposal | Examples & Templates. Scribbr. Retrieved February 16, 2024, from https://www.scribbr.com/research-process/research-proposal/
Is this article helpful?
Other students also liked, how to write a problem statement | guide & examples, writing strong research questions | criteria & examples, how to write a literature review | guide, examples, & templates, what is your plagiarism score.
Research School of Physics
- College of Science
- School structure
- Equity and diversity
- Annual reports
- ANU MakerSpace
- Physics outreach
- Research facilities
- Electronics Workshop
- Mechanical Workshop
- Physics endowments
- PhD and MPhil
- Intensive courses
- Find a research project
- Our student profiles
- Physics Education Centre
- Education outreach
- Summer Research Scholarships
- Project market day
- Physics data portal
- Explore Physics
- Find a project
- Do your PhD here
- Physics news
- Public events
- Meet our people
Potential student research projects
The Research School of Physics performs research at the cutting edge of a wide range of disciplines.
By undertaking your own research project at ANU you could open up an exciting career in science.
Student type All 3rdYear 1st year PhB later PhB Honours/MSc PhD/MPhil Summer Scholars
Department All Electronic Materials Engineering Fundamental & Theoretical Physics Materials Physics Nuclear Physics & Accelerator Applications Quantum Science & Technology Centre for Gravitational Astrophysics
Research field All Astrophysics Atomic and molecular physics Biophysics Clean energy Engineering in physics Environmental physics Fusion and plasma confinement Materials science and engineering Nanoscience and nanotechnology Photonics, lasers and nonlinear optics Physics of the nucleus Plasma applications and technology Quantum science and technology Theoretical physics Topological and structural science
Radioimpurities in particle detectors for dark matter studies.
Dr Michaela Froehlich , Dr Zuzana Slavkovska , A/Prof Stephen Tims , Professor Gregory Lane
Continuous gravitational waves from neutron stars
Distinguished Prof Susan Scott , Dr Lilli (Ling) Sun , Dr Karl Wette
Single atom counting for stellar nuclear synthesis studies
Dr Stefan Pavetich , Emeritus Professor Keith Fifield
Gravitational waves from ultralight boson clouds around black holes
Dr Lilli (Ling) Sun , Distinguished Prof Susan Scott , Dr Karl Wette
Gravitational waves from newborn neutron stars
Nanostructured Metasurfaces for Optical Telescopes
Dr Josephine Munro , Prof Andrey Sukhorukov
Paving the way to study the chronology of the early solar system
Dr Stefan Pavetich , Dr Michaela Froehlich , A/Prof Stephen Tims , Mr Dominik Koll
Calibration of gravitational wave detectors
Dr Lilli (Ling) Sun , A/Prof Bram Slagmolen , Distinguished Prof Susan Scott
Optimising a neutron star extreme matter observatory
A/Prof Bram Slagmolen , Dr Lilli (Ling) Sun , Distinguished Prof David McClelland
Multi-messenger gravitational-wave astronomy
Positron Annihilation Spectroscopy
Dr Joshua Machacek , Professor Stephen Buckman
Prospects of future ground-based gravitational-wave detector network
Dr Lilli (Ling) Sun , A/Prof Bram Slagmolen
How does a black hole ring?
Dr Lilli (Ling) Sun , Distinguished Prof Susan Scott
Exotic nuclear structure towards the neutron dripline
Dr AJ Mitchell
Atomic and Molecular Physics
Measuring and modelling free-ion hyperfine fields.
Professor Andrew Stuchbery , Emeritus Professor Tibor Kibedi , Dr Brendan McCormick
Positron applications in medical physics
A/Prof. James Sullivan , Professor Stephen Buckman , Dr Joshua Machacek
Interactions between antimatter and ultracold atoms
Dr Sean Hodgman , Professor Stephen Buckman , Dr Joshua Machacek
Atomic magnetometer for exploring physics beyond the standard model and gyroscopy
Professor Ben Buchler
Benchmark positron scattering experiments
Mass-entangled ultracold helium atoms
Dr Sean Hodgman , Professor Andrew Truscott
Positron interactions with structured surfaces
Dr Joshua Machacek , Dr Sergey Kruk
Electron and positron scattering from hydroxide, water and hydrogen peroxide
A/Prof. James Sullivan , Dr Edward Simpson
Optical quantum memory
Femtosecond laser for ultra-precise cavity drilling in modern dentistry
Dr Ludovic Rapp
Specific ion effects
Professor Vincent Craig
Solid-state nanopore sensors: Unveiling New Frontiers in Biomolecule Detection
Prof Patrick Kluth
Cross sections for nuclear fusion.
Dr Edward Simpson
Engineering in Physics
Nuclear structure studies with particle transfer reactions.
Dr AJ Mitchell , Professor Gregory Lane , Professor Andrew Stuchbery , Mr Ben Coombes
Nuclear lifetimes - developing new apparatus and methods
Professor Andrew Stuchbery , Emeritus Professor Tibor Kibedi , Professor Gregory Lane , Mr Ben Coombes
Wood-based mechanical metamaterials
Dr Nicolas Francois , Dr Mohammad Saadatfar , Professor Mark Knackstedt
Tracking noisy lasers using digitally enhanced fibre interferometers
Dr Chathura Bandutunga , A/Prof Malcolm Gray , Dr Paul Sibley , Dr Ya Zhang
Coherently combined laser systems for breakthrough starshot and beyond
Dr Chathura Bandutunga , Dr Paul Sibley , A/Prof Michael Ireland
Vibration control for optical interferometry
A/Prof Bram Slagmolen , Distinguished Prof David McClelland
Fibre optic sensor arrays for vibrometry and acoustic sensing
Dr Chathura Bandutunga , Dr Paul Sibley , A/Prof Malcolm Gray
Developing ultra-high resolution optical meta-surface sensors
Dr Chathura Bandutunga , Prof Dragomir Neshev
Directional dark matter measurements with CYGNUS
Dr Lindsey Bignell , Dr Peter McNamara , Dr Zuzana Slavkovska , Professor Gregory Lane
Miniature absolute gravimeter for long-term gravity surveys
Dr Samuel Legge , Professor John Close , Prof Patrick Kluth , Dr Giovanni Guccione
Ultra-fast lifetime measurements of nuclear excited states
Professor Gregory Lane , Dr AJ Mitchell , Professor Andrew Stuchbery , Emeritus Professor Tibor Kibedi
Understanding energy dissipation in colliding quantum many-body systems
Dr Kaitlin Cook , Dr Ian Carter , Professor Mahananda Dasgupta , Emeritus Professor David Hinde
Engineering Inter-spacecraft laser links
Professor Kirk McKenzie , Dr Andrew Wade
Total recall – memory effects in negative ion sources.
Radioactivity in our environment
Dr Michaela Froehlich
Surface forces and the behaviour of colloidal systems
High pressure non-equilibrium plasma discharges in chemically reactive systems
A/Prof Cormac Corr
Fusion and Plasma Confinement
Diagnosing plasma-surface interactions under fusion-relevant conditions.
A/Prof Cormac Corr , Dr Matt Thompson
Nano-bubble formation in fusion relevant materials
A/Prof Cormac Corr , Prof Patrick Kluth , Dr Matt Thompson
The effect of He irradiation on the microstructure and mechanical properties of W/ W alloys
Materials science and engineering, nanofluidic diodes: from biosensors to water treatment.
Neutron and X-ray imaging/tomography techniques at ANSTO & Australian Synchrotron
Dr Andrew Kingston , Dr Glenn Myers
Colloidal systems in highly concentrated salt solutions
Developing wearable sensors for personalized health care technologies and solutions
Dr Buddini Karawdeniya , Prof Dragomir Neshev , Prof Patrick Kluth , Professor Lan Fu
Ultra-low contact resistance next generation semiconductor devices
Emeritus Professor Robert Elliman , Mr Tom Ratcliff
Tomography of dynamic processes (3D movies)
Dr Andrew Kingston , Prof Adrian Sheppard , Dr Glenn Myers
Solid state synapses and neurons - memristive devices for neuromorphic computing
Emeritus Professor Robert Elliman , Dr Sanjoy Nandi
Machine learning for tomographic reconstruction
Dr Glenn Myers , Dr Andrew Kingston
Quantitative x-ray imaging with patterned illumination
Ultrafast laser cleaning - the light touch.
Spatial laser mode analysis for thermal noise measurements in optical cavities
Dr Johannes Eichholz , A/Prof Bram Slagmolen , Distinguished Prof David McClelland
High entropy alloys in advanced nuclear applications
A/Prof Cormac Corr , Dr Maryna Bilokur
Defect Engineering of 2D Materials
Emeritus Professor Robert Elliman
Optical nonlinearities in 2D crystals
Dr Giovanni Guccione , Professor Ping Koy Lam
Exciton polaritons in 2D atomically thin materials
Prof Elena Ostrovskaya , Professor Andrew Truscott
Ultrashort laser processing for advanced applications
Dr Ludovic Rapp , Professor Andrei Rode
Functional nanopore membranes
Creating new materials using pressure and diamond anvil cells
Prof Jodie Bradby
Nanowire photodetectors for photonic and quantum systems
Professor Lan Fu , Dr Ziyuan Li , Professor Hoe Tan
Shape engineering of semiconductor nanostructures for novel device applications
Professor Hoe Tan , Professor Chennupati Jagadish
High-bandwidth stabilisation of a 2µm-band laser
Measurement of optical and mechanical losses of mirror coatings
Deblur by defocus in a 3D X-ray microscope
Making diamond from disordered forms of carbon.
Can we make a new phase of carbon?
Efficient optical interconnect for quantum computers
Dr Rose Ahlefeldt
X-ray scatter in 3D microscopes
Dr Andrew Kingston , Dr Glenn Myers , Prof Adrian Sheppard
Nanoscience and Nanotechnology
Optical metamaterials: from science fiction to transformative optical technologies.
Prof Dragomir Neshev , Dr Andrei Komar , Dr Mohsen Rahmani
Nanowire infrared avalanche photodetectors towards single photon detection
Professor Lan Fu , Dr Zhe (Rex) Li , Professor Chennupati Jagadish
Micro-ring lasers for integrated silicon photonics
Metaphotonics and Mie-tronics with resonant dielectric structures
Professor Yuri Kivshar , Dr Kirill Koshelev
Engineering optical chirality with nanotechnology
Professor Yuri Kivshar , Dr Kirill Koshelev , Dr Sergey Kruk
Quantum-well nanowire light emitting devices
Professor Lan Fu , Dr Ziyuan Li , Professor Hoe Tan , Professor Chennupati Jagadish
Nanowire lasers for applications in nanophotonics
Professor Chennupati Jagadish , Professor Hoe Tan
Photonics, Lasers and Nonlinear Optics
Non-equilibrium quantum condensation of microcavity exciton polaritons.
Satellite based geodesy
Dr Syed Assad , Professor Ping Koy Lam , Mr Lorcan Conlon , Dr Jie Zhao
Synthesising non-Hermitian gauge fields for microcavity exciton polaritons
Dr Eliezer Estrecho , Prof Elena Ostrovskaya
Integrated quantum photonics
Prof Andrey Sukhorukov , Dr Jinyong Ma , Dr Jihua Zhang , Prof Dragomir Neshev
Nonlinear topological photonics
Dr Daria Smirnova
Low-noise offset-phase locking and heterodyne interferometry with 2µm-band lasers
Machine learning for optics and controls
A/Prof Bram Slagmolen
Laser levitation of a macroscopic mirror
Quantum squeezed states for interferometric gravitational-wave detectors
Distinguished Prof David McClelland , Professor Daniel Shaddock , A/Prof Bram Slagmolen
Quantum photonics with nanostructured metasurfaces
Dr Jinyong Ma , Prof Andrey Sukhorukov , Dr Jihua Zhang
Prof Dragomir Neshev , Prof Andrey Miroshnichenko
Synthetic multi-dimensional photonics
Prof Andrey Sukhorukov , Dr Jihua Zhang
Physics of the Nucleus
Nuclear magnetism - magnetic moment measurements.
Professor Andrew Stuchbery , Emeritus Professor Tibor Kibedi , Professor Gregory Lane , Dr Brendan McCormick
Nuclear batteries: Energy-storage applications of nuclear isomers
Dr AJ Mitchell , Professor Gregory Lane
Time dependence of nuclear fusion
Nuclei that fall apart: the role of sub-zeptosecond processes in reactions of weakly-bound nuclei
Dr Kaitlin Cook , Professor Mahananda Dasgupta , Emeritus Professor David Hinde
Towards a global understanding of nuclear fission
Dr Kaitlin Cook , Emeritus Professor David Hinde , Professor Mahananda Dasgupta
Nuclear vibrations in near-spherical and deformed nuclei
Professor Andrew Stuchbery , Professor Gregory Lane , Dr AJ Mitchell , Mr Ben Coombes
Plasma Applications and Technology
Quantum science and technology, quantum multi-parameter estimation.
Experimental quantum simulation with ultracold metastable Helium atoms in an optical lattice
Dual torsion pendulum for quantum noise limited sensing
Beam matching using machine learning
Dr Syed Assad , Dr Aaron Tranter , Dr Jie Zhao
Quantum super resolution
Dr Syed Assad , Professor Ping Koy Lam , Dr Jie Zhao
Variational approach to many-body problems.
A/Prof Vladimir Mangazeev
Introduction to quantum integrable systems
Stochastic dynamics of interacting systems and integrability
Combinatorics and integrable systems
A/Prof Vladimir Mangazeev , Professor Vladimir Bazhanov
Topological and Structural Science
Ghost imaging in the third dimension.
Video generation models as world simulators.
We explore large-scale training of generative models on video data. Specifically, we train text-conditional diffusion models jointly on videos and images of variable durations, resolutions and aspect ratios. We leverage a transformer architecture that operates on spacetime patches of video and image latent codes. Our largest model, Sora, is capable of generating a minute of high fidelity video. Our results suggest that scaling video generation models is a promising path towards building general purpose simulators of the physical world.
- View Sora overview
This technical report focuses on (1) our method for turning visual data of all types into a unified representation that enables large-scale training of generative models, and (2) qualitative evaluation of Sora’s capabilities and limitations. Model and implementation details are not included in this report.
Much prior work has studied generative modeling of video data using a variety of methods, including recurrent networks, [^1] [^2] [^3] generative adversarial networks, [^4] [^5] [^6] [^7] autoregressive transformers, [^8] [^9] and diffusion models. [^10] [^11] [^12] These works often focus on a narrow category of visual data, on shorter videos, or on videos of a fixed size. Sora is a generalist model of visual data—it can generate videos and images spanning diverse durations, aspect ratios and resolutions, up to a full minute of high definition video.
Turning visual data into patches
We take inspiration from large language models which acquire generalist capabilities by training on internet-scale data. [^13] [^14] The success of the LLM paradigm is enabled in part by the use of tokens that elegantly unify diverse modalities of text—code, math and various natural languages. In this work, we consider how generative models of visual data can inherit such benefits. Whereas LLMs have text tokens, Sora has visual patches . Patches have previously been shown to be an effective representation for models of visual data. [^15] [^16] [^17] [^18] We find that patches are a highly-scalable and effective representation for training generative models on diverse types of videos and images.
At a high level, we turn videos into patches by first compressing videos into a lower-dimensional latent space, [^19] and subsequently decomposing the representation into spacetime patches.
Video compression network
We train a network that reduces the dimensionality of visual data. [^20] This network takes raw video as input and outputs a latent representation that is compressed both temporally and spatially. Sora is trained on and subsequently generates videos within this compressed latent space. We also train a corresponding decoder model that maps generated latents back to pixel space.
Spacetime latent patches
Given a compressed input video, we extract a sequence of spacetime patches which act as transformer tokens. This scheme works for images too since images are just videos with a single frame. Our patch-based representation enables Sora to train on videos and images of variable resolutions, durations and aspect ratios. At inference time, we can control the size of generated videos by arranging randomly-initialized patches in an appropriately-sized grid.
Scaling transformers for video generation
Sora is a diffusion model [^21] [^22] [^23] [^24] [^25] ; given input noisy patches (and conditioning information like text prompts), it’s trained to predict the original “clean” patches. Importantly, Sora is a diffusion transformer . [^26] Transformers have demonstrated remarkable scaling properties across a variety of domains, including language modeling, [^13] [^14] computer vision, [^15] [^16] [^17] [^18] and image generation. [^27] [^28] [^29]
In this work, we find that diffusion transformers scale effectively as video models as well. Below, we show a comparison of video samples with fixed seeds and inputs as training progresses. Sample quality improves markedly as training compute increases.
Variable durations, resolutions, aspect ratios
Past approaches to image and video generation typically resize, crop or trim videos to a standard size—e.g., 4 second videos at 256x256 resolution. We find that instead training on data at its native size provides several benefits.
Sora can sample widescreen 1920x1080p videos, vertical 1080x1920 videos and everything inbetween. This lets Sora create content for different devices directly at their native aspect ratios. It also lets us quickly prototype content at lower sizes before generating at full resolution—all with the same model.
Improved framing and composition
We empirically find that training on videos at their native aspect ratios improves composition and framing. We compare Sora against a version of our model that crops all training videos to be square, which is common practice when training generative models. The model trained on square crops (left) sometimes generates videos where the subject is only partially in view. In comparison, videos from Sora (right) have improved framing.
Training text-to-video generation systems requires a large amount of videos with corresponding text captions. We apply the re-captioning technique introduced in DALL·E 3 [^30] to videos. We first train a highly descriptive captioner model and then use it to produce text captions for all videos in our training set. We find that training on highly descriptive video captions improves text fidelity as well as the overall quality of videos.
Similar to DALL·E 3, we also leverage GPT to turn short user prompts into longer detailed captions that are sent to the video model. This enables Sora to generate high quality videos that accurately follow user prompts.
Prompting with images and videos
All of the results above and in our landing page show text-to-video samples. But Sora can also be prompted with other inputs, such as pre-existing images or video. This capability enables Sora to perform a wide range of image and video editing tasks—creating perfectly looping video, animating static images, extending videos forwards or backwards in time, etc.
Animating DALL·E images
Sora is capable of generating videos provided an image and prompt as input. Below we show example videos generated based on DALL·E 2 [^31] and DALL·E 3 [^30] images.
Extending generated videos
Sora is also capable of extending videos, either forward or backward in time. Below are four videos that were all extended backward in time starting from a segment of a generated video. As a result, each of the four videos starts different from the others, yet all four videos lead to the same ending.
We can use this method to extend a video both forward and backward to produce a seamless infinite loop.
Diffusion models have enabled a plethora of methods for editing images and videos from text prompts. Below we apply one of these methods, SDEdit, [^32] to Sora. This technique enables Sora to transform the styles and environments of input videos zero-shot.
We can also use Sora to gradually interpolate between two input videos, creating seamless transitions between videos with entirely different subjects and scene compositions. In the examples below, the videos in the center interpolate between the corresponding videos on the left and right.
Image generation capabilities
Sora is also capable of generating images. We do this by arranging patches of Gaussian noise in a spatial grid with a temporal extent of one frame. The model can generate images of variable sizes—up to 2048x2048 resolution.
Emerging simulation capabilities
We find that video models exhibit a number of interesting emergent capabilities when trained at scale. These capabilities enable Sora to simulate some aspects of people, animals and environments from the physical world. These properties emerge without any explicit inductive biases for 3D, objects, etc.—they are purely phenomena of scale.
3D consistency. Sora can generate videos with dynamic camera motion. As the camera shifts and rotates, people and scene elements move consistently through three-dimensional space.
Long-range coherence and object permanence. A significant challenge for video generation systems has been maintaining temporal consistency when sampling long videos. We find that Sora is often, though not always, able to effectively model both short- and long-range dependencies. For example, our model can persist people, animals and objects even when they are occluded or leave the frame. Likewise, it can generate multiple shots of the same character in a single sample, maintaining their appearance throughout the video.
Interacting with the world. Sora can sometimes simulate actions that affect the state of the world in simple ways. For example, a painter can leave new strokes along a canvas that persist over time, or a man can eat a burger and leave bite marks.
Simulating digital worlds. Sora is also able to simulate artificial processes–one example is video games. Sora can simultaneously control the player in Minecraft with a basic policy while also rendering the world and its dynamics in high fidelity. These capabilities can be elicited zero-shot by prompting Sora with captions mentioning “Minecraft.”
These capabilities suggest that continued scaling of video models is a promising path towards the development of highly-capable simulators of the physical and digital world, and the objects, animals and people that live within them.
Sora currently exhibits numerous limitations as a simulator. For example, it does not accurately model the physics of many basic interactions, like glass shattering. Other interactions, like eating food, do not always yield correct changes in object state. We enumerate other common failure modes of the model—such as incoherencies that develop in long duration samples or spontaneous appearances of objects—in our landing page .
We believe the capabilities Sora has today demonstrate that continued scaling of video models is a promising path towards the development of capable simulators of the physical and digital world, and the objects, animals and people that live within them.
- Bill Peebles
- Connor Holmes
- David Schnurr
- Troy Luhman
- Eric Luhman
- Clarence Wing Yin Ng
- Aditya Ramesh
Please cite as Brooks, Peebles, et al., and use the following BibTeX for citation: https://openai.com/bibtex/videoworldsimulators2024.bib