Call for proposals
Project proposals should be a maximum of 4 pages (single column, minimum 10 pt font), plus 1 page for references, plus CV. All content components (proposal, references, CV) should be composed into a single PDF file. We accept one submission per Principal Investigator.
Project proposals must include:
- Motivation for a specific research task.
- An outline of the approach, challenges and expected results.
- A proposed budget not exceeding $80,000 in cash and $20,000 in AWS Promotional Credits.
- CV of the leading faculty member. CV can be submitted as a link listed at the end of the proposal document.
- Recognition: categorization, detection, segmentation
- Visual search
- Deep neural network compression and optimization
- Video understanding: actions, events
- Large-scale data annotation
- Computer vision for apparel
- Human body: detection, tracking, pose
- Motion: segmentation, tracking
- 3D modeling: structure-from-motion, slam, stereo and reconstruction
- Computational photography
- Computer vision for robotics
- Faces and gestures
- Image and video captioning
Fairness in artificial intelligence
- Transparency, explainability, and accountability in AI systems
- Theories of computational/algorithm fairness and factors that affect algorithmic trustworthiness
- Ethical decision-support and decision-making systems
- Detecting and ameliorating adverse biases in data and algorithms, and fairness-aware design of algorithms
- Metrics and methods for designing, piloting, and evaluating systems that mitigate against adverse biases and ensure fairness, including the use of human-machine collaboration and decision support
- Statistical methods for detecting bias in systems as they are operating
Knowledge management and data quality
- Data cleaning for machine learning
- Graph mining from knowledge graphs and user behaviors
- Knowledge embedding
- Knowledge extraction from unstructured and semi-structured data
- Knowledge verification
- Knowledge-based search
- Large-scale data alignment and integration
- Leveraging structured knowledge in deep learning and recommendation
- Quantitative and logical error detection
Machine learning algorithms and theory
- Active learning and data cleaning
- Data and resource efficient learning
- Deep learning and representation learning
- Fair, explicable and interpretable learning
- Transfer and meta-learning
- Online and continual learning
- Parallel and distributed Learning
- Robust and privacy preserving learning
- Reinforcement learning
Natural language processing
- Advances in neural MT for noisy and user-generated content
- Chatbots and dialogue systems
- Detection of inappropriate content
- Efficient training and fast inference for neural MT
- Context-aware MT
- Explainability in neural NLP methods
- Fact extraction, verification and trustworthiness in unstructured data
- Multitask and reinforcement learning for MT
- Named entity translation and transliteration
- NLP applications in search
- Question answering
- Text summarization
- Narrative understanding
- Common sense inference
- AI methods for online advertising
- Algorithmic marketing
- Large scale experimentation and testing
- Learning mechanisms
- Measurement of brand advertising
- Online algorithms for targeting, bidding and pricing
- Optimizing for long term objectives
- Prediction, forecasting and automated decision making in ad systems
- Structure of advertising marketplaces
Operations research and optimization
- Assortment management
- Management of warehouse operations
- Marketplace design: incentives/policies for increasing efficiency and growth in a multi-agent marketplace
- Strategic supply chain management: network design/topology
- Tactical supply chain management: vendor management (including supplier contract negotiation and procurement), inventory buying, inventory deployment, demand fulfillment
- Transportation: long-haul operations (including airline operations), last-mile operations
- Other supply chain optimization topic
- Approaches to estimate quality of recommenders using abundant implicit and sparse explicit feedback
- Detecting and responding to spam in behavioral data to protect customers in recommendation contexts
- Scalable NLP approaches for search query understanding for non-English
- Scalable approaches to detect incorrect catalog information
- Approaches to identity synonyms in noisy product catalog
- Item-to-item collaborative filtering using deep learning
- Affective and social interactions
- Autonomous navigation and mobility
- Dexterous and reactive manipulation
- Human machine interaction and collaboration
- Machine learning and learning from human preferences
- Motion planning
- Multi-robot systems and multi-agent pathfinding
- Object detection and pose estimation
- Sample-efficient reinforcement learning
- Semantic scene understanding for robotics
- Simulation and sim to real transfer
- SLAM and long-term autonomy
- Theoretical advances as well as practical applications
Search and information retrieval
- Multilingual language understanding
- Conversational search
Security, privacy and abuse prevention
- ML for malware analysis and detection
- Browser/device fingerprint and digital forensics
- Early detection of emerging patterns with limited labeled data (one-shot-learning)
- Fraud detection and prevention
- Graph modeling (latent representations from a graph and anomaly detection)
- Human-in-the-loop machine learning
- Online and adaptive machine learning
- Web behavioral modeling, online identity and password-less authentication
- Threat and intrusion detection for cloud security
- ML for obfuscation detection from text, image and online behaviors
- Detection and tracking of online adversarial attempts
Last updated: September 4, 2019
By applying to or participating in the Amazon Research Awards Program (“ARA Program”), each member of the research team agrees to the following rules of engagement. The lead researcher/s who applies to the ARA Program (the “Principal Investigator/s”) is/are responsible for distributing these rules to all members of the research team before their participation in any research in connection with the proposal funded by the ARA Program. Please read these items carefully to understand the conditions of your participation in the ARA Program and receipt of ARA funding.
Eligibility to apply for the ARA Program depends on the following: The “Principal Investigator/s”) must: (1) be a permanent faculty member at an accredited college(s) or university(ies) (excluding colleges or universities located in any of the restricted jurisdictions outlined below) that grant PhD degrees in fields related to Machine Learning, (2) not have any other paid employment outside of academia, (3) be at or above the age of majority in their jurisdiction of residence at the time of application, and (4) not be a person under US export controls or sanctions. The ARA Program is void in Cuba, Iran, Syria, North Korea, Sudan, the region of Crimea, and where otherwise prohibited by law. Individuals are ineligible if they are, or have been within the last 12 months at the date of application to the ARA Program, directors, officers, employees, interns and contractors (“Personnel”) of Amazon.com Services Inc, and its Affiliates (“Amazon”), immediate family members of ineligible Personnel (parents, siblings, spouses, children), and members of the households of ineligible Personnel. Principal Investigators may only submit one proposal to the ARA Program per call for proposal period, however research team members may contribute to research under more than one proposal. For the purposes of these rules of engagement, “Affiliates” mean any entity that directly or indirectly controls, is controlled by, or is under common control with Amazon.com Services, Inc. Amazon reserves the right to verify the Principal Investigator and research team’s eligibility before it distributes ARA funding.
No proposal to the ARA Program may contain any confidential information and no part may be marked as ‘confidential’. Amazon does not accept any legal obligation, whether of confidentiality, compensation, return or otherwise, with respect to any proposals. You understand and acknowledge that Amazon has wide access to technology, designs, and other materials, and may work on and/or develop projects and ideas that may be competitive with, similar to, or identical to your proposal in theme, idea, format or other respects, inclusive. You acknowledge and agree that you will not be entitled to any compensation as a result of Amazon’s use of any such similar or identical material that has or may come to Amazon from other sources. Amazon reserves the right to implement competitive, similar, or identical ideas in the future, without restriction or obligation.
Your Own Work
You represent and warrant that your proposal (i) is the original work of you and your research team (or an update to original work); and (ii) does not, (1) to your knowledge, infringe any third party patent rights or (2) infringe, misappropriate or otherwise violate any other third party intellectual property rights (i.e., other than patent rights), including any copyrights, trade secrets, trademarks, contract or licensing rights, rights of publicity or privacy, or moral rights.
Amazon Research Awards (“ARAs”) are structured as a one-year unrestricted gifts to academic institutions. ARA funding amounts will be determined by Amazon in its sole discretion. ARA funding is not extendable or transferable, but you may submit new proposals for subsequent ARA Program calls. Except where prohibited by law, Amazon is not responsible for any tax deductions or withholdings which may be necessary under all federal, state, provincial or other tax liabilities, except that Amazon may withhold taxes as required by law. You or your academic institution are responsible for any costs and expenses associated with ARA funding acceptance and use. Amazon does not pay overhead or indirect costs on ARA funding.
AWS Promotional Credits
If ARA funding includes AWS Promotional Credits, the AWS Promotional Credit Terms and Conditions apply. These credits are provided for company/institutional use only; they are not for personal use. You should confirm there are no federal, state, local, or institutional ethics or procurement laws, regulations, or other rules that would restrict or prohibit your institution's receipt of and use of these credits, or would otherwise create a conflict of interest for AWS or its affiliates. For example, that there are no procurements/acquisitions for which your receipt and use of these credits could conflict AWS or its affiliates from participating in the procurement/acquisition. You should consult your ethics or procurement official in making this determination; please reach out to Amazon with any questions you may have.
Privacy and Marketing
You acknowledge and agree that Amazon may collect, store, share, and otherwise use personally identifiable information provided during the ARA application process, including but not limited to, name, mailing address, phone number, and email address. All personally identifiable information collected is subject to, and will be used in accordance with, Amazon’s Privacy Notice (available at www.amazon.com/privacy), including for administering ARA and verifying applicant’s identities, addresses, and telephone numbers in the event a proposal is selected by Amazon. By participating in the ARA Program, you consent to the transfer of personal data to the United States for purposes of administering the ARA Program, conducting publicity about the ARA Program and additional purposes consistent with Amazon’s goals relating to the ARA Program. The data controller for information collected by Amazon is Amazon.com Services, Inc, 410 Terry Ave North, Seattle, Washington 98109, USA. Except where prohibited, you consent to Amazon’s use of your names, likeness, biographical information, and voice in advertising, publicity, trade, and other marketing and promotional materials (including video, audio, and print through all means of distribution) worldwide without compensation, notice or approval, and you disclaim any ownership rights to the content of such materials.
Limitation of liability
EACH PRINCIPAL INVESTIGATOR AND RESEARCH TEAM MEMBER ACCEPTS THE CONDITIONS STATED IN THESE OFFICIAL RULES, AGREES TO BE BOUND BY THE DECISIONS OF AMAZON, WARRANTS THAT HE OR SHE IS ELIGIBLE TO PARTICIPATE. EACH RESEARCH TEAM MEMBER, THE PRINCIPAL INVESTIGATOR AND THE PRINCIPAL INVESTIGATOR’S INSTITUTION HEREBY RELEASES AMAZON FROM, AND WAIVES ANY AND ALL CLAIMS AGAINST AMAZON FOR, ANY LOSSES, LIABILITY, AND DAMAGES OF ANY KIND, (INCLUDING FOR ANY LOSS OF DATA, LOST PROFITS, COST OF COVER OR OTHER SPECIAL, INCIDENAL, CONSEQUENTIAL, INDIRECT, PUNITIVE, EXEMPLARY OR RELIANCE DAMAGES) INCURRED OR SUSTAINED IN CONNECTION WITH OR RISING OUT OF (1) THE ARA PROGRAM OR ANY TRAVEL OR ACTIVITY RELATED THERETO, (2) USE OF ANY PROPOSAL OR RIGHTS THEREIN, OR (3) ANY BREACH OF ANY AGREEMENT OR WARRANTY ASSOCIATED WITH THE ARA PROGRAM, INCLUDING THESE OFFICIAL RULES, HOWEVER CAUSED AND REGARDLESS OF THEORY OF LIABILITY.
We may amend any of these rules at our sole discretion by posting the revised terms on the ARA Program website. Your continued participation in the ARA Program after the effective date of the revised rules constitutes your acceptance of the rules.
Any dispute or claim relating in any way to your participation in the ARA Program will be resolved by binding arbitration, rather than in court, except that you may assert claims in small claims court if your claims qualify. The Federal Arbitration Act and federal arbitration law apply to this agreement.
There is no judge or jury in arbitration, and court review of an arbitration award is limited. However, an arbitrator can award on an individual basis the same damages and relief as a court (including injunctive and declaratory relief or statutory damages), and must follow these rules as a court would.
To begin an arbitration proceeding, you must send a letter requesting arbitration and describing your claim to our registered agent Corporation Service Company, 300 Deschutes Way SW, Suite 304, Tumwater, WA 98501. The arbitration will be conducted by the American Arbitration Association (AAA) under its rules, including the AAA's Supplementary Procedures for Consumer-Related Disputes. The AAA's rules are available at www.adr.org or by calling 1-800-778-7879. Payment of all filing, administration and arbitrator fees will be governed by the AAA's rules. We will reimburse those fees for claims totaling less than $10,000 unless the arbitrator determines the claims are frivolous. Likewise, Amazon will not seek attorneys' fees and costs in arbitration unless the arbitrator determines the claims are frivolous. You may choose to have the arbitration conducted by telephone, based on written submissions, or in person in the county where you live or at another mutually agreed location.
We each agree that any dispute resolution proceedings will be conducted only on an individual basis and not in a class, consolidated or representative action. If for any reason a claim proceeds in court rather than in arbitration we each waive any right to a jury trial. We also both agree that you or we may bring suit in court to enjoin infringement or other misuse of intellectual property rights.
All issues and questions concerning the construction, validity, interpretation and enforceability of these rules, or the rights and obligations of you or Amazon in connection with the ARA Program shall be governed by, and construed in accordance with, the laws of the State of Washington without giving effect to any choice of law or conflict of law rules (whether of the State of Washington or any other jurisdiction), which would cause the application of the laws of any jurisdiction other than the State of Washington.