Mozilla Research Grants are a program to help us keep the Internet safe, open, and accessible to all, as it evolves. Where appropriate, we are particularly looking for proposals that support our aim to grow an internet that truly puts people first, where individuals can shape their own experience and are empowered, safe and independent.
For the 2019H1 funding series, we will only be accepting proposals to answer 12 specific research questions, in three categories: Growing the Web, New Interaction Modes, and Privacy & Security. These are the research questions:
Growing the Web
The WASI standard enables WebAssembly runtimes to give selective access to resources and functionality of the host environment. E.g., it enables giving access to only a single file or directory instead of just giving free access to the system’s files. How and when the user is asked to give or deny access to specific resources is up to the WASI runtime. We are looking for someone to define, verify and ideally build an exemplar of the user experience for this system for our Wasmtime WASI runtime.
Rust uses a compilation technique called monomorphization, which generates specialized code for every instantiation of a generic function. While this is crucial for performance, it also results in code bloat. Can we characterize the extent to which this code bloat is responsible for Rust's compilation times? If we were to enable a hybrid compilation strategy that sometimes monomorphized and sometimes did not (e.g., via a dictionary-passing compilation), what kinds of improvements to compilation time are possible? We are looking for someone to build a prototype tool (perhaps based on rustc itself) that can estimate what improvements are possible for a given source program.
RQ3: Digital trust for cost-conscious users
After conducting several field research studies in Southeast Asia, particularly Indonesia, we are seeing a particular pattern in cost-conscious users’ mobile behavior around digital trust. Digital trust is a critical factor when people shop online through e-commerce site or social media, especially in emerging markets where there may be few or no trusted platforms to support digital trust between people. How can a browser or a website responsibly increase digital trust on the internet, and what are the particular opportunities for these emerging markets?
RQ4: Beyond OBA
Online behavioral advertising (OBA), where digital advertisers use personal data to target users, has become the predominant funding model for the web. OBA practices can help consumers find products they are more likely to be interested in, but tracking online behavior and sharing this personal data raises privacy concerns. We are interested in projects that investigate viable alternative funding models for the web that will increase user privacy and autonomy.
New Interaction Modes
RQ5: Iodide: R or Julia in the Browser
Fathom is a supervised-learning system for recognizing parts of web pages. How can we receive corrections from users of Fathom applications without violating their privacy? Federated learning, differential privacy, or dimensionality reduction might be pieces of a solution. What's the best tradeoff you can reach between the collection of practical corrective (and validative) value and preservation of privacy?
As internet-enabled technologies move from screens into our ears, eyes, and embedded public spaces, we will face a growing thicket of cultural and intergenerational conflicts in values around privacy, control, and trust. How might we better understand, acknowledge, address, and mediate conflicts in values in the context of building and evolving software products? How might we better understand how software products impact social values and vice versa? We’re particularly interested in specific cases that help us understand these impacts where there are significant imbalances of power and privilege.
RQ8: Mixed Reality
We are interested in understanding how the web creates new or unique AR/VR opportunities for consumers. How can social VR and AR best be integrated with live events and remote participation? What are new ways for consumers to create and share AR and VR content and experiences on the web? How can speech be integrated with web-based mobile AR applications to make it easier to interact with 3D content through a 2D interface?
Current voice interfaces are generally triggered by the use of a wakeword, although many people don’t understand the privacy and security implications of such devices. What would a respectful, privacy-conscious, and, above all, useful model of interacting with a continuously listening, non-keyword-triggered voice interface look like? What would need to be true for people to trust the system enough to use it, and what would it be used for?
RQ10: Common Voice
What defines a “quality” voice dataset in any given language for training machine learning voice technology? This is not only the number of hours, number of speakers, and demographics, but also understanding the techniques for classifying the data, working with accents, and labeling the regions. How do we get better at collecting crowd-sourced voice data? What techniques have we yet to try? What are the metrics of success, and how might we enhance processes to create improved voice datasets?
Privacy & Security
A substantial amount of social interactions, information-seeking behavior, and productivity happen online through the browser. In Firefox we can run experiments that allow us to collect data and design interventions in order to examine, explore, and investigate these areas of inquiry. However, Mozilla has an ongoing strong interest in enabling user control over data collection efforts and respecting our users’ agency in data collection initiatives. We are interested in proposals that seek not only to understand how people interact with the Web via study of browsing data (and how the Web responds to their behaviors, such as targeted advertising or descriptive investigations of algorithmic bias), but also to understand how people understand and interact with consent architectures as they relate to participation in online experiments, data collection requests, and other methods of large-scale, passive data collection.
RQ12: Privacy & Security for Firefox
Mozilla has an interest in potentially integrating more of Tor into Firefox, for the purposes of providing a Super Private Browsing (SPB) mode for our users. Tor offers privacy and anonymity on the Web, features which are sorely needed in the modern era of mass surveillance, tracking and fingerprinting. However, enabling a large number of additional users to make use of the Tor network requires solving for inefficiencies currently present in Tor so as to make the protocol optimal to deploy at scale. Academic research is just getting started with regards to investigating alternative protocol architectures and route selection protocols, such as Tor-over-QUIC, employing DTLS, and Walking Onions. What alternative protocol architectures and route selection protocols would offer acceptable gains in Tor performance? And would they preserve Tor properties? Is it truly possible to deploy Tor at scale? And what would the full integration of Tor and Firefox look like?
We absolutely recognize the value of research outside of these domains and research questions, but for this round we will only accept funding for work in these categories.
Our aim with this program is to support Mozilla’s research efforts, rather than to generally fund research on the internet. In the absence of a strong Mozilla champion and/or a strong tie to current research, historically there are a few approaches we’re unlikely to fund:
We generally don’t fund the development of new open source projects through Mozilla Research Grants except for cases where we specifically request as such. We are looking to fund the creation of new research knowledge in the world, and in some cases this may involve releasing open source code, but that can't be the primary aim of the proposal.
We generally don’t fund proposals to improve the state of closed ecosystems, such as app stores.
We generally don’t fund studies of particular populations using or not using technologies, unless those are tied to very particular Mozilla goals.
We don’t fund development of ongoing open source (or closed source) programs, except as specified in the research questions: such programs should go through MOSS.
Your first step should be to read through the whole submission form, well in advance, so you know what you're going to need to do. We expect you to put together a clear, coherent expression of your problem, the approach you're taking to solve that problem, and the solution you hope to achieve. You might want to think of this as a business plan for your research effort. What's the minimum level of success you're hoping to achieve and what's the best case? How will you know how successful you are? How many people will be impacted by what you do if you’re successful? What would happen if you didn’t do this research?
Make sure that someone who isn't an expert in your field can understand your project, and make sure that someone who is an expert in your field can understand what you're doing that's new.
You should make sure that you explicitly cover these criteria: viability, alignment, value, and impact.
viability: how likely is it that you will be able to execute on what you plan?
alignment: how well is this project aligned with the specific request?
value: is this project good value for money?
impact: how much will this project impact the world? how many people will it impact? what would happen if you don't do this project?
Incorporating a plan for ongoing engagement with Mozilla is a great plus, as we’ve found the most successful projects are those with strong ongoing interactions. If there are Mozilla data sources you would like to collaborate around, such as Firefox, Common Voice, or survey data, please specify that in the proposal. We’ve found successful projects often have an existing internal champion; if you don’t know who the right champion for your work at Mozilla might be, first complete as much of your proposal as you can, and then email us, by May 23rd, and we will attempt to find you an internal champion.
What limitations are there for this funding?
Applications must be affiliated with a university, research institute or research-focused registered non-profit, in any country except for those embargoed by the US State Department. Gifts are awarded to those institutions; we cannot award them to unaffiliated individuals. You must include a plan for disseminating the results, which would normally include publication in a peer-reviewed and open-access venue, and we encourage you to make those publications, results, code, and/or data publicly accessible. We will pay open-access fees for not-for-profit publishers included in your budget. Please acknowledge Mozilla's support in your publication, and send it to us when it gets published. We particularly encourage you to further publish your work in a format more accessible to the public, like blog posts or articles in the popular press. Send us those, too! We will ask for roughly two status updates a year.
University-affiliated applicants can be students or faculty; students will require a letter from their advisor. Funding amounts are set at $25,000 including all expenses. We expect the timescale for most projects, not counting final publications, to be around one year, although that is only a guideline. We strongly encourage ongoing collaboration with Mozilla over the entire course of the grant.
As part of our commitment to diversity, we will fund childcare up to 10% of a grant, with a cap of $2500. We particularly encourage applications from new faculty in their first or second years. Funding is given as an unrestricted gift to the institution. We do not pay university overhead.
Please make sure you've checked the FAQ before you start writing your application; please ask questions through that same page. In addition, we will have live online office hours in this Vidyo room on Monday, 20 May at 9:00am PDT, and Tuesday, 21 May at 6:00pm PDT.
The submission deadline is Friday, May 31th at 2:22pm Pacific Daylight Time (PDT). The next deadline will be announced in the Fall of 2019; please sign up for our mailing list if you wish to be notified.
When you're ready to begin, hit the big Start button at the top right.