Apple Scholars in AI/ML Fellowship
Destefani, Bel (Deactivated)
Giulia
Omer, Betty
Apple's 2024 Scholars in AI/ML: https://machinelearning.apple.com/updates/apple-scholars-aiml-2024
2025-26 Cohort Call for Nominations
UC San Diego has been invited to submit nominations for the Apple Scholars in AI/ML PhD Fellowship. The Apple Scholars in AI/ML PhD fellowship program recognizes the contributions of emerging leaders in computer science and engineering at the graduate and postgraduate level. The PhD fellowship in AI/ML was created as part of the Apple Scholars program to support the work of outstanding PhD students from around the world, who are pursuing cutting edge research in machine learning and artificial intelligence.
Award Details
The fellowship award is comprised of a generous monetary gift, mentorship, and potential internship opportunities. All monetary gifts are directed towards the nominating university and disbursed annually at the beginning of each academic year, and are conditional upon the Scholar’s full time enrollment in their program.
- Gift amount covering full tuition and fees (enrollment fees, health insurance) for (2) academic years
- $40,000 USD stipend each year to help with living expenses and related expenses
- $5,000 USD stipend each year to support research-related travel and associated expenses
- Mentorship with an Apple researcher
- Internship opportunities during their fellowship*
*Internship offers are dependent on student status, and contingent upon necessary requirements for employment being met according to relevant employment law.
Nominees must meet the following criteria to be considered:
- Nominee must be enrolled full time at the nominating university at the start of Fall 2025, and expect to be enrolled through the end of the 2025/2026 academic year
- Nominee should be entering their last 2-3 years of study as of Fall 2024
- Nominee must not hold another industry-sponsored full fellowship while they are an Apple Scholar in AI/ML (Fall 2025 to Summer 2027)
Apple believes that technology for everyone should be made by everyone, and that research is strengthened by a diversity of perspectives and lived experiences. They aim to create an inclusive and equitable nomination process that amplifies underrepresented voices in the research community. With that in mind, invited schools may nominate students per the following guidelines.
Nomination Rules
- UCSD may nominate a total of (3) PhD students pursuing research relevant to the research areas listed below. Each department may nominate two (2) students. In the event that we get more than two (2) nominations from campus departments we will return the nominations to the department to submit only the top two (2) students. The Graduate Fellowship Review Committee will decide which three (3) students will become the final nominees to represent UC San Diego.
- Each nominated student must be submitted for a unique research area. Apple will not be accepting nominations for multiple students under a single research area.
- Departments are strongly encouraged by Apple to use at least (1) of the slots to nominate students who identify as a member of a traditionally underrepresented group in the technology industry.
- An individual nominee’s underrepresented group status will not be collected in the application by Apple, and applications will be reviewed based solely on the strength of the submitted materials.
- Students may not self-nominate and faculty members must work with their department Graduate Coordinator to formally nominate their students.
What is an underrepresented group, and how is it defined?
An underrepresented group is typically defined as a group whose representation in a particular context is significantly lower than their group size in the wider population. In the US, the term underrepresented minority (URM) is used, and in the US technology industry it generally refers to those who identify as Black, Hispanic, Native American, Native Hawaiian, or Pacific Islander. Other underrepresented groups in the technology industry include women and non-binary individuals.
Underrepresented groups or communities may be defined differently by other industries and in other locations, and can change over time.
The final decision is left up to the school and its nominees to interpret and define “traditionally underrepresented group” within their own regional and cultural context.
Nomination Materials
Universities should submit the following materials for each student’s nomination:
- Student CV and publication list
- Research Abstract (200 word maximum)
- Research statement covering past work and proposed direction for next 2 years (5 page maximum, including citations, in a legible font size) clearly stating the hypothesis and expected contributions to the chosen research area.
- 2 letter of recommendation, one from current advisor (1 page maximum per letter)
- Link to most recent published work (optional)
- Documents must be submitted as one (1) PDF with the file naming convention:
- LastName_FirstName_PID_AppleFellowship_AY2425
Nominees should be pursuing research in one or more of the following research areas. The subtopics listed under each research area are not meant to be exhaustive or prescriptive, but rather highlight areas of particular interest to Apple. When entering a nomination, the nominator will be asked to select up to (2) research areas that the nominee feels are most aligned with their work. There is no single research area that Apple prioritizes over another. Schools and nominees are encouraged to select the area(s) most relevant to the research.
Research Area | Description |
---|---|
Privacy Preserving Machine Learning | Privacy Preserving Machine Learning focuses on developing techniques to analyze data without compromising the privacy of individuals. At Apple, we believe that privacy is a human right. The goal of this area is to enable the sharing of data while ensuring that sensitive information remains protected. Sub topics: federated learning, differential privacy, cryptographic tools, secure multiparty computation |
Human Centered AI | Human Centered AI seeks to design, develop, and deploy AI systems that are aligned with human values and needs. Emphasizing the importance of involving human perspectives, feedback, and insights throughout the AI development process to ensure that the technology is beneficial to society. Sub topics: social signal processing, ML for multimodal interaction, ML design and human factors, usable ML tools and products, interactive ML, model personalization, human-in-the-loop ML |
AI for Ethics and Fairness | AI for Ethics and Fairness focuses on developing AI systems that are unbiased and ethical. Seeking to address issues such as algorithmic bias, discrimination, and transparency, in order to ensure that AI is used in a fair and just manner. Sub topics: bias and fairness in AI, interpretable AI, introspection, robustness |
AI for Accessibility | AI for Accessibility focuses on developing AI systems to help people with disabilities to interact with the world in new ways. Emphasizes the importance of creating inclusive technology that enables equal access and participation for all. Sub topics: accessible user experiences, automatic personalization/adaptation, interactions via new or combined modalities, participatory design with people with disabilities |
AI for Health and Wellness | AI for Health and Wellness on developing AI systems to improve healthcare outcomes and promote personal wellness. This involves the use of machine learning, statistics, and signal processing to analyze health data, support decision-making, and recommend personalized care. Sub topics: ML and RL for mobile health, time series representation learning, physiology informed machine learning, modeling multi-modal sensor data, causal modeling, human behavior |
ML Theory | ML Theory focusses on understanding the mathematical foundations and theoretical properties of machine learning algorithms. ML Theory seeks to explain how and why different algorithms work, and to identify the limits of what can be learned from data. Sub topics: understanding ML, generalization, physics-based ML, generative models, imbalanced data theory, out-of-distribution setting |
ML Algorithms and Architectures | ML Algorithms and Architectures focuses on developing new algorithms, models, and architectures to improve the performance and efficiency of machine learning. Sub topics: auto ML, model compression, architecture / search, optimization, model representation, interpretability, large-scale ML, foundation models, imbalanced data, unsupervised and self supervised representation learning, diffusion models |
Embodied ML | Embodied ML focuses on developing intelligent agents that can learn to interact with the physical world through trial-and-error learning. Involving the integration of reinforcement learning with robotics, computer vision, and natural language processing to create agents that can perceive, reason, and act in the real world. Sub topics: imitation learning, multi-output models, reinforcement learning for embodied ML, hardware/software integration, hardware aware ML training, inference and resource constrained ML |
Speech and Natural Language | Speech and Natural Language focuses on developing algorithms and models to understand or generate spoken or written human language. This involve the use of statistical modeling and machine learning techniques such as deep learning to build systems that can interpret and respond naturally to human language. Sub topics: speech recognition, text to speech, conversational and multi-modal interactions, machine translation, language modeling and generation, large language models |
Computer Vision | Computer Vision is a field that focuses on developing algorithms and models to analyze and interpret visual information captured through digital interfaces. Techniques such as deep learning using convolutional neural networks and transformers enable the development of powerful systems that can recognize, classify, and semantically interpret visual data. Sub topics: semantic scene understanding, video understanding , 3D scene understanding , efficient deep learning for computer vision, AI for content creation, continual learning, computer vision for AR/VR, computer vision with synthetic data, language and vision, computational photography, vision for robotics, foundation model for industrial machine vision, vision for industrial robotics |
Information Retrieval, Ranking and Knowledge | Information Retrieval and Knowledge focuses on developing algorithms and models to extract, organize and infer meaningful information from large amounts of data and serve such information to satisfy user needs. |
Data-Centric AI | Data-Centric AI focuses on developing machine learning techniques to effectively manage, process, and analyze large volumes of data, while also ensuring data efficacy, efficiency, and fairness. This includes the creation and generation of high-quality datasets, through techniques such as synthetic data generation, data augmentation, and active learning, as well as the annotation and curation of data for machine learning models. Sub topics: data efficacy, data efficiency, data generation, data fairness, synthetic data generation, dataset creation, data and annotation, active learning, ML-enabled data annotation, augmentation and curation, transfer learning with limited data, multi modal language models, unsupervised and weakly-supervised anomaly detection, Synthetic defect generation and simulation, sim2real transfer learning |
The Apple AI/ML Fellowship is a gift under unique Fund TBD. GEPA's Interim Dean, Judy Kim, serves as the PI on this award.
Since this is a gift, please work with GEPA's Financial Analyst and UCSD's Gift Processing Office to complete the appropriate award acceptance documents.
This fellowship should be used as per the intent of the Apple Fellowship written in the award letter.
Chart of Accounts:
- Each awardee will receive a unique project tied to their financial unit under unique Fund TBD.
- Once the project has been established, please request appropriate AIDID's in order to use the fellowship