OVERVIEW

As part of one of the world's most innovative and recognizable brands, we are committed to support university research and innovation in the U.S., Canada, India, and select European countries, while also fostering partnerships with university faculty and researchers. The Sony Research Award Program provides funding for cutting-edge academic research and helps build a collaborative relationship between faculty and Sony researchers. With awards up to $150,000 USD* per year for each accepted proposal, both the Faculty Innovation Award and Focused Research Award create new opportunities for university faculties and research institutions to engage in pioneering research that could drive new technologies, industries and the future.

AWARD RECIPIENTS

Congratulations to all award recipients in the Sony Research Award Program! We sincerely look forward to working closely with you.

2024 (more renewals pending)

• Professor Gopala Anumanchipalli, University of California, Berkeley, United States
• Professor Gedas Bertasius, University of North Carolina at Chapel Hill, United States
• Professor Pushpak Bhattacharya, Indian Institute of Technology Bombay, India
• Professor Filippo Maria Fazi, University of Southampton, United Kingdom
• Professor Ioannis Gkioulekas, Carnegie Mellon University, United States
• Professor Boqing Gong, Boston University, United States
• Professor Natasha Jaques, University of Washington, United States
• Professor Chenfanfu Jiang, University of California, Los Angeles, United States
• Professor Vaibhav Katewa, Indian Institute of Science, India
• Professor Hun-Seok Kim, University of Michigan, United States
• Professor Phillip Koehn, Johns Hopkins University, United States
• Professor Karen Liu, Stanford University, United States
• University of Texas at Austin, United States
• Professor Owen Miller, Yale University, United States
• Professor Ramesh Raskar, Massachusetts Institute of Technology, United States
• Professor Craig Shultz, University of Illinois Urbana-Champaign, United States
• Professor Haoning Tang, University of California, Berkeley, United States
• Professor Andrea Vedaldi, University of Oxford, United Kingdom
• Professor Lei Wang, Lancaster University, United Kingdom
• Professor Ming-Hsuan Yang, University of California, Merced, United States
• Professor Kwang Moo Yi, University of British Columbia, Canada
• Professor Zhi-Li Zhang, University of Minnesota Twin Cities, United States

2023 (more renewals pending)

• Professor Haitham Al Hassanieh, EPFL, Switzerland
• Professor Harry Atwater, California Institute of Technology, United States
• Professor Changhyun Choi, University of Minnesota Twin Cities, United States
• Professor Jia Deng, Princeton University, United States
• Professor Rafael Gómez-Bombarelli, Massachusetts Institute of Technology, United States
• Professor Laurie Heller, Carnegie Mellon University, United States
• Professor Marco Hutter, ETH Zurich, Switzerland
• Professor Ranjay Krishna, University of Washington, United States
• Professor Johannes Levin, Ludwig-Maximilians-Universität München, Germany
• Professor Yunzhu Li, Columbia University, United States
• Professor David Lindell, University of Toronto, Canada
• Professor David Lindell, University of Toronto, Canada***
• Professor Xuezhe Ma, University of Southern California, United States
• Professor Alan Marshall, University of Liverpool, United Kingdom
• Professor Euan McLeod, University of Arizona, United States
• Professor Pratyush Mishra, University of Pennsylvania, United States
• Professor Ken Nakagaki, University of Chicago, United States
• Professor Sang-Hyun Oh, University of Minnesota Twin Cities, United States
• Professor Allison Okamura, Stanford University, United States
• Professor Hayder Radha, Michigan State University, United States
• Professor Alan Ritter, Georgia Institute of Technology, United States
• Professor Qi Sun, New York University, United States
• Professor Molei Tao, Georgia Institute of Technology, United States
• Professor Chen Wang, University at Buffalo, United States
• Professor Nozomu Yachie, University of British Columbia, Canada
• Professor Junsong Yuan, University at Buffalo, United States
• Professor Yuke Zhu, University of Texas at Austin, United States
• Professor Mohit Gupta, University of Wisconsin-Madison, United States**
• Professor Chenfanfu Jiang, University of California, Los Angeles, United States**
• Professor Ewold Verhagen, NWO Institute AMOLF, The Netherlands**
• Professor Bolei Zhou, University of California, Los Angeles, United States**

2022

• Professor Mario Barbatti, Aix-Marseille Université, France
• Professor Gedas Bertasius, University of North Carolina at Chapel Hill, United States
• Professor Debashis Chanda, University of Central Florida, United States
• Professor Angel Chang, Simon Fraser University, Canada
• Professor Zaijun Chen, University of Southern California, United States
• Professor Gregory Durrett, University of Texas at Austin, United States
• Professor Jaime Fernandez Fisac, Princeton University, United States
• Professor Chuang Gan, University of Massachusetts Amherst, United States
• Professor Tatsunori Hashimoto, Stanford University, United States
• Professor Chenfanfu Jiang, University of California, Los Angeles, United States
• Professor Howard Lee, University of California, Irvine, United States
• Professor Yong Jae Lee, University of Wisconsin-Madison, United States
• Professor Yaping Liu, Northwestern University, United States
• Professor Psyche Loui, Northeastern University, United States
• Professor Michele Magno, ETH Zurich, Switzerland
• Professor Alan Marshall, University of Liverpool, United Kingdom
• Professor Julian McAuley, University of California, San Diego, United States
• Professor Prineha Narang, University of California, Los Angeles, United States
• Professor Dimitrios Nikolopoulos, Virginia Polytechnic Institute and State University, United States
• Professor Sarah Ostadabbas, Northeastern University, United States
• Professor Shinsuke Shimojo, California Institute of Technology, United States
• Professor Bradley Siwick, McGill University, Canada
• Professor Aneta Stefanovska, Lancaster University, United Kingdom
• Professor Lili Su, Northeastern University, United States
• Professor Ewold Verhagen, NWO Institute AMOLF, The Netherlands
• Professor Nandita Vijaykumar, University of Toronto, Canada
• Professor Yu Xiang, University of Texas at Dallas, United States
• Professor Bolei Zhou, University of California, Los Angeles, United States
• Professor Pulkit Agrawal, Massachusetts Institute of Technology, United States**
• Professor Federico Capasso, Harvard University, United States**
• Professor Aaron Courville, University of Montreal, Canada**
• Professor Mohit Gupta, University of Wisconsin-Madison, United States**
• Professor Aiichiro Nakano, University of Southern California, United States**
• Professor Jose Principe, University of Florida, United States**
• Professor Zhou Yu, Columbia University, United States**

2021

• Professor Michal Bajcsy, University of Waterloo, Canada
• Professor Gedas Bertasius, University of North Carolina at Chapel Hill, United States
• Professor Federico Capasso, Harvard University, United States
• Professor Li-Jing Cheng, Oregon State University, United States
• Professor Aaron Courville, University of Montreal, Canada
• Professor Jia Deng, Princeton University, United States
• Professor Guillermo Gallego, Technische Universität Berlin, Germany
• Professor Aditya Grover, University of California, Los Angeles, United States
• Professor Charles Hages, University of Florida, United States
• Professor Cho-Jui Hsieh, University of California, Los Angeles, United States
• Professor Dongyeop Kang, University of Minnesota Twin Cities, United States
• Professor Pan Li, Georgia Institute of Technology, United States
• Professor Jun Liu, University at Buffalo, United States
• Professor Saturnino Luz, The University of Edinburgh, United Kingdom
• Professor Arka Majumdar, University of Washington, United States
• Professor Stephen Morris, University of Oxford, United Kingdom
• Professor Aiichiro Nakano, University of Southern California, United States
• Professor Andrew Owens, University of Michigan, United States
• Professor Srijith P.K, Indian Institute of Technology Hyderabad, India
• Professor Bryan Pardo, Northwestern University, United States
• Professor Arpita Patra, Indian Institute of Science, India
• Professor Jose Principe, University of Florida, United States
• Professor Ravi Ramamoorthi, University of California, San Diego, United States
• Professor Arindam Sanyal, Arizona State University, United States
• Professor Sebastian Scherer, Carnegie Mellon University, United States
• Professor Alireza Vahid, University of Colorado Denver, United States
• Professor Sheng Wang, University of Washington, United States
• Professor Xiaolong Wang, University of California, San Diego, United States
• Professor Benjamin Williams, University of California, Los Angeles, United States
• Professor Lei Zhou, University of Texas at Austin, United States
• Professor Pulkit Agrawal, Massachusetts Institute of Technology, United States**
• Professor Mohit Gupta, University of Wisconsin-Madison, United States**
• Professor Chia Wei Hsu, University of Southern California, United States**
• Professor Deepak Pathak, Carnegie Mellon University, United States**
• Professor Zhou Yu, Columbia University, United States**

2020

• Professor Jacob Andreas, Massachusetts Institute of Technology, United States
• Professor Veselka Boeva, Blekinge Institute of Technology, Sweden
• Professor Changhyun Choi, University of Minnesota Twin Cities, United States
• Professor Steven Cummer, Duke University, United States
• Professor Guido de Croon, Delft University of Technology, The Netherlands
• Professor Krzysztof Gajos, Harvard University, United States
• Professor Rafael Gómez-Bombarelli, Massachusetts Institute of Technology, United States
• Professor Aman Haque, The Pennsylvania State University, United States
• Professor Felix Heide, Princeton University, United States
• Professor Chia Wei Hsu, University of Southern California, United States
• Professor Kyle Jamieson, Princeton University, United States
• Professor Evangelos Kalogerakis, University of Massachusetts Amherst, United States
• Professor Patanjali Kambhampati, McGill University, Canada
• Professor Bhaskar Krishnamachari, University of Southern California, United States
• Professor Fenglong Ma, The Pennsylvania State University, United States
• Professor Shuji Nakamura, University of California Santa Barbara, United States
• Professor VP Nguyen, University of Texas at Arlington, United States
• Professor Deepak Pathak, Carnegie Mellon University, United States
• Professor Marc Pollefeys, ETH Zurich, Switzerland
• Professor Fang Song, Portland State University, United States
• Professor Natalie Stingelin, Georgia Institute of Technology, United States
• Professor Gordon Wetzstein, Stanford University, United States
• Professor Joerg Wrachtrup, University of Stuttgart, Germany
• Professor Jun Yao, University of Massachusetts Amherst, United States
• Professor Junming Yin, Carnegie Mellon University, United States
• Professor Pulkit Agrawal, Massachusetts Institute of Technology, United States**
• Professor David Bishop, Boston University, United States**
• Professor Mark Brongersma, Stanford University, United States**
• Professor Siyang Cao, University of Arizona, United States**
• Professor Virginia de Sa, University of California, San Diego, United States**
• Professor Mohit Gupta, University of Wisconsin-Madison, United States**
• Professor Po-Chun Hsu, Duke University, United States**
• Professor Pedro Lopes, University of Chicago, United States**
• Professor Dimitris Papailiopoulos, University of Wisconsin-Madison, United States**
• Professor Daniel Sanchez, Massachusetts Institute of Technology, United States**
• Professor Aswin Sankaranarayanan, Carnegie Mellon University, United States**
• Professor Zhou Yu, Columbia University, United States**
• Professor Jun-Yan Zhu, Carnegie Mellon University, United States**

2019

• Professor Alan Aspuru-Guzik, University of Toronto
• Professor Farrokh Ayazi, Georgia Institute of Technology
• Professor Yoshua Bengio, University of Montreal
• Professor Mark Brongersma, Stanford University
• Professor Siyang Cao, University of Arizona
• Professor Virginia de Sa, University of California, San Diego
• Professor Jonathan Fan, Stanford University
• Professor Kenan Gundogdu, North Carolina State University
• Professor Mohit Gupta, University of Wisconsin-Madison
• Professor Po-Chun Hsu, Duke University
• Professor Yong Jae Lee, University of California, Davis
• Professor Jian Lin, University of Missouri
• Professor Pedro Lopes, University of Chicago
• Professor Xiaoping Qian, University of Wisconsin-Madison
• Professor Xiang Ren, University of Southern California
• Professor Aswin Sankaranarayanan, Carnegie Mellon University
• Professor Akane Sano, Rice University
• Professor Randy Sweis, University of Chicago Medicine
• Professor Zhou Yu, University of California, Davis
• Professor Xinyu Zhang, University of California, San Diego
• Professor Jun-Yan Zhu, Carnegie Mellon University
• Professor Pulkit Agrawal, Massachusetts Institute of Technology**
• Professor David Bishop, Boston University**
• Professor Oliver Cossairt, Northwestern University**
• Professor Song Han, Massachusetts Institute of Technology**
• Professor Xingjie Ni, The Pennsylvania State University**
• Professor Dimitris Papailiopoulos, University of Wisconsin-Madison**
• Professor Daniel Sanchez, Massachusetts Institute of Technology**
• Professor Ted Sargent, University of Toronto**

2018

• Professor Pulkit Agrawal, Massachusetts Institute of Technology
• Professor David Bishop, Boston University
• Professor Oliver Cossairt, Northwestern University
• Professor Jonathan Fan, Stanford University
• Professor Katerina Fragkiadaki, Carnegie Mellon University
• Professor Bolin Liao, University of California Santa Barbara
• Professor Patrick Lin, Cal Poly, San Luis Obispo
• Professor Wojciech Matusik, Massachusetts Institute of Technology
• Professor Florian Metze, Carnegie Mellon University
• Professor Xingjie Ni, The Pennsylvania State University
• Professor Dimitris Papailiopoulos, University of Wisconsin-Madison
• Professor Alexander Rush, Cornell University
• Professor Ted Sargent, University of Toronto
• Professor Sebastian Scherer, Carnegie Mellon University
• Professor Muhammad Shahzad, North Carolina State University
• Professor Todd Sulchek, Georgia Institute of Technology
• Professor Shimeng Yu, Georgia Institute of Technology
• Professor John Zhang, Dartmouth College
• Professor Song Han, Massachusetts Institute of Technology**
• Professor Aggelos Katsaggelos, Northwestern University**
• Professor Daniel Sanchez, Massachusetts Institute of Technology**

2017

• Professor Faramarz Fekri, Georgia Institute of Technology
• Professor Kristen Grauman, University of Texas at Austin
• Professor Song Han, Massachusetts Institute of Technology
• Professor Daniel Sanchez, Massachusetts Institute of Technology
• Professor Xinyu Zhang, University of California, San Diego
• Professor Stefano Ermon, Stanford University**
• Professor Aggelos Katsaggelos, Northwestern University**
• Professor Shivendra Panwar, New York University**

2016

• Professor Dirk Bernhardt-Walther, University of Toronto
• Professor Stefano Ermon, Stanford University
• Professor Aggelos Katsaggelos, Northwestern University
• Professor Scott Kuindersma, Harvard University
• Professor Bruno Olshausen, University of California, Berkeley
• Professor Shivendra Panwar, New York University

• *Awards will be paid in USD or the respective currency depending on the country of the participating university/institution.
• **Renewed Research Collaboration
• ***Principal investigator won an award for each of different research topics.

FACULTY INNOVATION AWARD

Global research and development at Sony enables us to foster innovative ideas, which could ultimately lead to future technology advancements and company growth. In order to speed up and expand the creation of new ideas, we would like to partner with universities and research institutes. This partnership will help cultivate advanced concepts and fertilize our own research and development. The Sony Faculty Innovation Award provides up to $100K USD* in funds to conduct pioneering research in the areas listed below. Please select the single most relevant keyword to your submission. In an effort to further connect with users and creators alike, for many of the keywords listed below, creator tools or creator technologies are often one of the envisioned use-cases.

Information Technology

Keywords are bulleted under each category title

Entertainment Technology

Keywords are bulleted under each category title

Device Technology

Keywords are bulleted under each category title

• *Awards will be paid in USD or the respective currency depending on the country of the participating university/institution.
• **Brain Machine Interface.

FOCUSED RESEARCH AWARD

Solid research is the underlying driving force to crystallize fearless creativity and innovation. While we are committed to run in-house research and engineering, we are also excited to collaborate with academic partners to facilitate exploration of new and promising research. The Sony Focused Research Award provides an opportunity for university faculty, research institutes, and Sony to conduct this type of collaborative, focused research. The award provides up to $150K USD* in funds, and may be renewed for subsequent year(s). A list of candidate research topics appears below. Please select the Focused Research Theme for which your submission is written.

Visual Chain-of-Thought and Simplified Responses Using Visual Elements

In the entertainment industry, conducting quality checks and risk assessments of content before delivery is essential, while ensuring time and cost efficiency to keep the project on schedule. AI assistance is expected to significantly enhance the efficiency of these processes. However, the current Vision-Language Models (VLM) face the following challenges:

• Detailed Analysis in Time and Space: It is difficult to conduct fine-grained analyses along timelines or pixel spaces.
• Iterative Analysis from Multiple Perspectives: There is a lack of ability to scrutinize content repeatedly from diverse viewpoints.
• Visual and Intuitive Output: It is challenging to explain the analysis results of visual content with text-only outputs.

Scope of Proposal:

To address these challenges, Sony is seeking the development of innovative analytical methods such as, but not limited to the following:
Utilizing Visual Chain-of-Thought and External Knowledge/Tools

• Automatically identify temporal and spatial areas within the target content that require attention,
• Collaborate with image manipulation (e.g., zoom-in) and external tools (e.g., detection models) to actively repeat visual reasoning by highlighting areas deemed necessary for analysis,
• Verify the consistency between the content and external documents (such as specifications and guidelines, which may include visual instructions), multiple times and from diverse perspectives,

VLM Outputting Responses Incorporating Visual Information

• Generate responses that users can intuitively understand by combining not only text information but also visual elements (e.g., bounding boxes, arrows, highlights, etc.), and
• Output operations of drawing tools to explain visually.

Multimodal Retrieval-Augmented Generation for Creative Assistance

Sony is seeking and developing technologies that support creative industries, such as filmmaking and music production. Professional creative workflows are inherently multimodal, combining language, vision, and sound. Recent advances in AI have improved the ability to align information across these modalities, enabling more efficient utilization of assets such as visual storyboards, soundtrack references, and narrative templates in creative workflows. Retrieval-Augmented Generation (RAG) systems can enhance these workflows by grounding the processes in reliable and explainable materials. However, while RAG has shown promising performance in text-based applications, extending it to multimodal data remains a significant challenge as the model needs to understand and reason over the queries, intentions, and documents across multiple modalities. We are looking for novel technologies of multimodal RAG that integrate language, vision, sound and beyond, to better support creative workflows.

Scope of Proposal:

Novel technologies for creative assistance, including but not limited to the following:

• Multimodal-LLMs and multimodal reasoning for creative assistance,
• Multimodal embedding for better RAG,
• Generative retrieval for RAG in multiple modalities,
• Efficient and scalable approaches to RAG, and
• New benchmarks and evaluation methods for multimodal RAG.

Next-Gen Communication Paradigms: Semantic and Task/Goal-oriented Communication

As the demand for intelligent, efficient, and context-aware communication systems continues to grow, semantic and task/goal-oriented communication has emerged as a transformative paradigm. Unlike traditional communication models that focus solely on bit-level accuracy, these approaches aim to convey the intended meaning of information and facilitate the achievement of specific tasks and goals. By prioritizing meaning and purpose, these novel technologies enable more efficient and intelligent interactions between machines and humans.

Scope of Proposal:

Topics of interest include, but are not limited to the following;

• Investigate and establish a comprehensive theoretical framework for semantic and task/goal-oriented communication, including semantic information metrics, goal-driven transmission models, and task-aligned communication strategies,
• Explore and design core algorithms for semantic and goal-oriented encoding, decoding, and inference, leveraging advances in machine learning, natural language processing, and signal processing,
• Define system-level models that bridge theory and application, identifying key components and architectures for real-world deployment of meaning-aware and goal-aware communication systems,
• Develop and validate proof-of-concept (PoC) systems, demonstrating the feasibility and effectiveness of semantic and task/goal-oriented communication in practical scenarios, and
• Conduct experimental studies and simulations to evaluate performance, robustness, and scalability of proposed models and algorithms for meaningful and goal-driven communication.

Physics-based Robotics Simulation

Physics-based simulation is one of the most important technologies that accelerates robotics for manipulation and locomotion. It is also applicable to generate more realistic motion of virtual characters for entertainment. There has been great recent progress made in complex simulation to reproduce authentic environments, such as accurate real-time deformation, photo-realistic rendering, and AI/ML for simulation. However, many challenges remain. For example, dexterous manipulation requires physically-accurate, various-object simulations (including soft, rigid, fluid, fragile, etc. objects), and methods on how to fill the gap between the real world and virtual environment are crucial. Finer and larger-scale complex objects, and various biological physical behaviors/operations have not been achieved in simulation so far. Sony is interested in novel simulation technologies for robotics and entertainment.

Scope of Proposal:

Innovative simulation technologies for robotics may include but are not limited to the following:

A fast, accurate, large-scale, and complex objects simulation

• 10x faster physics simulation method/algorithm,
• Various objects simulation (rigid, soft, fluid, fine, etc.),
• ML accelerated physics simulation and physics-informed NN for robotics,
• Efficient parallel processing to speed-up large-scale simulations,
• Handling contact between robots, environment, and objects accurately,

Technologies that enhance a loop between real and virtual worlds

• Novel Sim2Real/Real2Sim,
• Multimodal sensor simulation,
• Text to motion for dynamic, complex physical interaction, and
• Semantics simulation.

Incorporating New Optical Functions to High-peak-power Solid-state Surface Emitting Laser

Sony has realized a high-peak-power chip-scale surface-emitting laser*(WL-DPSSEL: Wafer-Level Diode-pumped Solid-state Surface-Emitting Laser) that produces pulses as short as 450 ps and an estimated peak power of 57.0 kW with a laser chip volume of 1 mm³. We plan to further enhance its performance with cutting-edge technology through this program.
*reference

• Chip-scale high-peak-power semiconductor/solid-state vertically integrated laser | Nature Communications

Scope of Proposal:

We seek novel technologies to enable our high-peak-power solid-state surface-emitting lasers to evolve further in various areas such as, but not limited to, the following:

• Incorporation of an acousto-optic, electro-optic, or magneto-optic modulator into a solid-state surface-emitting laser to enable dynamic control of the Q-factor and laser beam properties including pulse-to-pulse timing stability, polarization, longitudinal mode, etc.,
• Incorporation of a meta-surface into solid-state surface emitting lasers to enable novel optical functions, and
• Incorporation of functional layers into solid-state surface emitting lasers to enable novel optical functions.

Innovative Visual Technology powered by AI

Recent advances in machine learning have created a paradigm shift for many applications. For instance, deep learning-based approaches have achieved a big leap forward over the previous state-of-the-art in segmentation, recognition, and reconstruction. These approaches, including generative AI, continue to evolve. Sony is looking for innovative research in image/video processing based on machine learning to significantly improve existing image/video processing techniques and applications in 3D/4D as well as 2D. Sony's goal and interest in this area is to create new value for creators through entirely new products or services.

Scope of Proposal:

Topics of interests include:

• Generative AI such as image to photo-realistic video, video to video, style transfer, modal transfer based on new approach, e.g. photo-realistic image generation using neural rendering or novel generative model with high controllability,
• Multimodal content generation based on visual language models (VLMs),
• Image/video compression for viewing and sensing such as neural representation, deep learning hashing, generative coding,
• Multi-view image/video generation and 3D/4D model generation/reconstruction, e.g., novel view synthesis using neural rendering, neural inverse rendering such as NeRF and 3D/4D gaussian splatting,
• Perceptual metrics for predicting photo-realistic image quality or 3D image quality and determining physical inconsistencies in motion from photorealistic videos or 4D data,
• Low-latency processing and complexity/computational cost reduction for the above applications, and
• Training data creation such as CG utilization and generative AI for the above applications.

AI-based Digital Human Content Creation

Sony is looking for innovative research in data-driven novel (imaginary but photo-realistic and believable) digital human or character creation techniques based on AI and machine learning.
The goal is to significantly improve existing hand-crafted visual art production workflow and to develop a new set of tools to complement Sony creators'creativity.

Scope of Proposal:

Topics of interests include:

• AI-based novel digital human avatar or photorealistic character creation from 2D image, video, audio/speech or text, and
• AI-based novel facial/body performance/motion synthesis without motion capture (expression, emotional speech), natural simulation (hair, clothes) or behavior synthesis (reaction, interaction) from 2D image, video, audio/speech or text.

Understanding Gameplay and Game Scenes through Multimodal LLMs

In recent years, Multimodal Large Language Models, including LLMs and VLMs, have attracted considerable attention. We invite research proposals leveraging recent advances in foundation models (LLMs, VLMs, multimodal AI) to analyze and enhance player experience, performance, and agent behavior in video games. Two thematic tracks are outlined below.

1. Gameplay and Game Scene Understanding from Gameplay Footage
   We invite research proposals for methods that effectively utilize such foundation models to 1) understand the mechanics, rules, and context of each player's gameplay from gameplay videos, 2) store such information as external knowledge, 3) use the stored knowledge for subsequent analysis and evaluation of gameplay videos. Here, "understanding context" goes beyond simple summarization. For instance, it is essential that a VLM that doesn't know about a particular game, be able to watch a tutorial presented as gameplay footage, and acquire an understanding of the game's mechanics like a human, and then utilize that knowledge to interpret subsequent gameplay. We envision that such techniques will contribute to the realization of virtual agents that are capable of providing context-sensitive assistance and coaching from outside of the game environment.
2. Rewards for Gameplay
   We invite research proposals for developing dynamic methods for identifying and extracting meaningful gameplay events, specifically for applications in autonomous game-playing agents and AI-driven companion systems. For instance, consider a shooter game where a player successfully dodges incoming enemy fire. This can be recognized as a significant gameplay event. Such an event can be used in two major ways. The first way is as reinforcement for autonomous agents. These events can act as feedback or rewards in training autonomous agents, helping them learn effective strategies through reinforcement learning. For example, consistently avoiding enemy attacks could be positively reinforced to train the agent to prioritize evasive actions. The second way is for conversation with AI companions. An AI gaming companion can utilize these detected events to provide insightful, personalized commentary to the player. If a player asks, "How am I doing?" the AI can respond based on a contextual understanding of performance metrics such as successful dodges, hit accuracy, or objective completion, rather than just raw statistics.

   • To achieve this, research could leverage recent advances in foundational models and focus on scalable algorithms to (a) automatically discover which in-game events are relevant or noteworthy; (b) link those events to game-specific objectives (e.g., surviving, scoring, capturing); (c) or relate the identified events to a player's overall performance and behavior over time.
   • This requires combining game telemetry, contextual game state awareness, and methods to scale for game genres and titles to ensure that feedback or agent behavior generalizes well and that it is grounded in meaningful and actionable game data.

Scope of Proposal:

Sony seeks proposals on one of the sub-topics under either of the two major themes listed below.

1. Gameplay and Game Scene Understanding from Gameplay Footage
   • Gameplay and game scene understanding from gameplay footage,
   • Acquisition, storage, and utilization of dynamic knowledge from gameplay footage, or
   • Gameplay Evaluation (e.g., Skill Level, Proficiency/Deficiency)
2. Rewards for Gameplay
   • Dynamic approaches to identify game events from game videos that include game,audio, audio, play video and on-screen text,
   • Learning algorithms to link game events to the overall game play goal and strategyfor both individual levels and the entire game title,
   • Automatically discover game play correlates to user's game play performance, or
   • Focus on large data, scalable, single and multi-modal ML approaches.

Encoding and Protocols for Distributed Computing

A foundational technology for Cloud-Native Gaming (CNG) is the ability to run real-time interactive graphical applications and experiences across multiple devices in a cloud environment. This is what would take a feature from cloud-hosted to truly cloud-native, and allow for cloud tech to go beyond console limitations.
Latency and bandwidth are a major hurdle for having GPUs across different devices achieve this, due to the time-critical, real-time need for GPU processes to stay in sync.
Novel network protocols, emerging interconnects, and data center fabrics offer the potential to scale accelerators and facilitate rapid communication between them.
Novel ways to treat data streaming, neural encoding, and scene encoding specific to games and 3D scenes also offer potential in this area, as well as possibly enabling new methods of interacting with cloud development environments or understanding streamed information.

Scope of Proposal:

The goal of this proposal is to explore novel technologies and protocols in the area of network interconnects, accelerators, neural encoding, data center design, and GPU-to-GPU communication. We are particularly interested in how they could impact the quality of a game being streamed from a data center to a consumer. These are some of the research questions to be explored:

• How can a graphical or game experience be scaled across more devices in a data center, while retaining the quality and responsiveness that would be expected of a local game? How can we maintain performance or guarantee a base level of performance?
• Can neural or scene encoding be used to enhance specific aspects of a streamed game, or to improve the streaming quality in ways specific to a 3D scene? How does this change how a developer interacts with their game in development? What new information or viewpoints into a scene does this enable?
• Are there novel ways to orchestrate data centers specifically for interactive and game experiences?

Design and Evaluation of an Intermediate Interaction Layer for Integrating Diverse Input/Output Environments

Recent advances in multimodal sensing, such as gaze, voice, gesture, and biometric signals have significantly changed the way user inputs and intentions are interpreted in generative AI technologies.
In this context, there is a growing need for abstraction-layer technologies capable of inferring a user's operational intent from arbitrary input/output devices and physiological states, and of automatically mapping them to application commands and feedback modalities in an adaptive manner.
Furthermore, context-aware, dynamically generated feedback including visual, auditory, and haptic responses using AI technologies alongside translation and optimization of nonverbal communication in asynchronous settings, has emerged as a critical area of interest.

Scope of Proposal:

Sony seeks proposals for research on high-level abstraction, automatic transformation, and adaptive control technologies that enable seamless interaction, expression, and shared experience across diverse environments, users, and devices, namely;

• Research on integrating multimodal inputs (e.g., gaze, voice, gesture, biometric data) to infer user intent and interaction context,
• Design and implementation of intermediate abstraction layers or command representations that translate arbitrary input/output devices into unified control schemes,
• Development of adaptive feedback mechanisms that dynamically generate and select appropriate responses (visual, auditory, haptic, etc.) based on user state and environmental context,
• Techniques for supporting multimodal communication across heterogeneous device environments, including translation of nonverbal cues (facial expression, posture, prosody, etc.), and
• Prototyping and evaluation of the integrated technologies in real-world applications such as games, XR environments, and remote communication platforms.

Flexible Structural Mechanisms Enabling Dynamic Motion Entertainment

There are many humanoid and animal-mimicking robots in use.
Many pet robots, such as companion robots, are modeled after various animals, while robots expected to replace humans in various manufacturing settings, as well as service robots, are predominantly humanoid, regardless of their mobility capabilities.
For pet robots, service robots, and entertainment robots that require interaction with humans, the primary objective is not to perform tasks such as moving or assembling objects, but rather the act of "moving" itself, which necessitates expressive movement.
The expressiveness of robots designed for entertainment purposes is determined by their ability to mimic the movements of their reference subjects and their flexibility; however, existing robots often have a limited number of joints, which, while sufficient for performing specific tasks, fall short in terms of expressiveness.
Many existing robots, both humanoid and animal-mimicking, predominantly imitate vertebrates but are constructed with a limited number of joints. In particular, the joints in the torso typically possess only 0 to 3 degrees of freedom, resulting in a disparity between their movements and those of the subjects they imitate, significantly diminishing their entertainment value.
While some robots achieve diverse expressiveness through numerous joints and external power sources, they lack mobility and are designed for fixed operation.
Robots that can move autonomously and possess flexibility have the potential to offer unprecedented location-based entertainment (LBE) and consumer services that make movement itself a form of entertainment.

Scope of Proposal:

• Sony seeks proposals that will enable it to realize a small robot mimicking vertebrates, including humans, with a body length of approximately 30 cm.
• The system and technology will include structures, actuators, and control methods that realize flexible body movements akin to those of animals.
• The design will achieve a large range of motion while maintaining soft movements.
• While the inclusion of a mobility mechanism is not necessary, all essential elements must be contained within the robot, given the premise of mobility.

• *Awards will be paid in USD or the respective currency depending on the country of the participating university/institution.

SUBMISSION GUIDELINES

Eligibility

Eligible Applicants

We accept applications from Principal Investigators (PIs) who meet the eligibility criteria below.

• PI's institution: PIs must belong to a university, educational institution, or governmental/non-profit research institute. (E.g. we accept submissions from research institutes that are categorized as governmental and NPO/NGO in nature index tables.)
• PI's position: PIs must be a full-time professor (adjunct professors and adjunct researchers are not eligible) or researcher and be eligible to supervise Ph.D. students at the PI's institution. Full professors, associate professors, and assistant professors are eligible to apply. For the UK, this includes Senior Lecturers, Principal Lecturers, Lecturers, and Readers who are eligible to supervise Ph.D. students at their institutions.
• Countries: The PI's institution must be in Austria, Belgium, Canada, Denmark, Finland, France, Germany, India, Ireland, Italy, Luxembourg, the Netherlands, Norway, Poland, Portugal, Spain, Sweden, Switzerland, the United Kingdom, and the United States of America.

Co-PI(s)

We accept applications that have a PI and one or more co-PIs for the same proposal as long as all co-PIs are from the same institution as the PI. However, only one award is made to the PI and the PI's university/institution if the proposal is selected. All co-PIs must meet the same eligibility criteria as that for a PI, and co-PIs will be required to sign program documents.

Multiple Proposal Submissions

A PI or different PIs from the same university/institution may submit more than one proposal for different research topics. However, we ask each PI to please not submit identical proposals, and for each PI to not submit more than one proposal for each research topic. Sony will gladly accept an unlimited number of proposals from the same university for the same research topic as long as they each have different PIs. Note that at the bottom of your submission confirmation email, there is a link to resubmit your proposal in the event that you discover that your original had an error or an omission. Please use this link for resubmissions. Do not resubmit identical proposals. Also, please do not submit "test" submissions. In the rare event that you do not receive a confirmation email for your submission, please complete the Inquiry Form available here informing us of your issue.

Requirements

Target Award

Please select one target award between the Focused Research Award and the Faculty Innovation Award when you submit a proposal. Refer to the following comparison chart:

Target Research Topic

Please select one Focused Research Theme for your proposal if you choose the Focused Research Award. Please select one keyword (the single most relevant keyword for your proposal) if you choose the Faculty Innovation Award.

Research Agreement

A sponsored research agreement is required that is mutually agreed upon by Sony**, the PI, and his/her institution on all program terms including objectives, milestones, publication, use of research, and patent rights before any award is made or funding is available. This sponsored research agreement is negotiable at the award stage and must be between Sony**, the PI, and the PI's university/institution only. The sponsored research agreement will be for the benefit of the Sony legal entity specifically contracting with the PI's university/institution as well as Sony's corporate affiliates.

Term of Proposals

Please submit proposals for a one-year period only. No multi-year proposals will be accepted. An extension of the research may be possible depending upon the results of the research collaboration in year one, but it will require a separate discussion for another award the following year.

Minimum Deliverables

Three (3) quarterly reports and a final research summary report are required at a minimum.

Visiting Researcher Support

Sony may request the PI to support a Sony visiting researcher(s) at Sony's option and expense.

Confidentiality

Proposal authors or universities/institutions must ensure that no confidential or proprietary information is included in submitted proposals. Sony will treat all information submitted in proposals as non-confidential and non-proprietary.

Intellectual Property (IP) Rights

IP rights are negotiable at the award stage and will be specified in the sponsored research agreement. At a minimum, Sony and its corporate affiliates require the right to utilize the results of the research that Sony or its corporate affiliates sponsor for noncommercial purposes, including results that describe potentially patentable subject matter. Sony and its corporate affiliates will also require rights to any background IP that is required to utilize the research results sponsored. No submissions are allowed that use background IP to which the PI and his/her institution do not have the full authority to grant noncommercial use rights to Sony and its corporate affiliates.

Proposal Format

Please include the required items listed below in your submission:

• Title, abstract, methods, goals/milestones, references, and either one Focused Research Theme (if for the Focused Research Award) or one primary keyword (if for the Faculty Innovation Award).
• Please describe your submission's differentiation from the current state-of-the-art.
• Please include the best contact email address and phone number (complete with country code) for the PI.

All proposal contents must fit within 11 pages (a ten-page maximum proposal with references and a one-page budget summary). The file format must be a PDF file format and must be under 16 MB in size. Out of consideration to reviewers, please limit your minimum font size to a 10-point font.

Curriculum Vitae (CV)

The CV for the PI must be included when you submit a proposal. The CV file must be a separate file from the proposal file. There is no page limitation for the CV, however the CV file size must be under 16 MB.

Budget Guidelines

The Focused Research Award is limited to a maximum of $150K USD* per proposal. The Faculty Innovation Award is limited to a maximum of $100K USD* per proposal. This funding is a sponsored research grant that is to be used to conduct the research described in the proposal, and includes any overhead related to this research and any and all other fees or charges needed to carry out the research. There is a single payment for the award and it is all-inclusive of all associated expenses and fees. Sony will not specify a maximum amount or percentage of the budget that may be allocated to overhead.

Submission

Submission Terms and Conditions

The submitter must agree to the 2025 Sony Research Award Program Submission Terms and Conditions (click here to see) before submitting a proposal. Please obtain a prior review of your submission by your institution's sponsored research office if they require a review of research proposals prior to their submission.

Submission Protocol

Submissions must be done through the online submission form.

• Click here to open the submission form

Submissions made by email will not be accepted. Submissions must be in the English language or they will not be accepted. Submitters will automatically receive a confirmation email once they complete the online submission form and process. Please keep the confirmation email as proof of each submission.

Deadline

Submissions must be completed by 11:59 pm PDT (Pacific Daylight Time; UTC-7) on September 15, 2025 / 8:59 am CEST (Central European Summer Time; UTC+2) September 16, 2025 / 12:29 pm IST (Indian Standard Time; UTC+05:30) September 16, 2025.

Duplicate Submissions and Resubmissions

Duplicate submissions will not be accepted for the same proposal title from the same PI. Please do not make duplicate submissions. Use the resubmission link at the bottom of your confirmation email if you discover that your submission has an error or an omission.

Required Information for the Online Submission Form

The following information will be required in order to complete the online submission form for each proposal submitted.

• Proposal information; proposal title, proposal file (11 pages maximum including proposal, references, and 1-page budget summary by research quarter), and either one Focused Research Theme (if for the Focused Research Award) or one primary keyword (if for the Faculty Innovation Award).
• PI's Information; PI's full name, PI's title, PI's CV file, PI's email address, and PI's department name.
• Only for the US and Canada, please enter your SAT International Code for PI's university/institute (please find PI's SAT International code here). Please be careful to use the correct code because some schools have several codes to designate different programs or campuses. Also note that some school names use common abbreviations (e.g., INST for INSTITUTE) or simply truncate some words in the school's name when searching for a school code. Please enter "0000" if you cannot find the four-digit SAT International Code for your school.
• Submitter's Information, if not the PI (optional); Submitter's full name and email address

Award Candidates

Announcement of Review Results

Principal Investigators (PIs) will be notified of the submission review results for each proposal submitted around March of 2026. Sony will only be able to provide some limited feedback on the review process results for the highest-ranked proposals due to limited resources.

Funding

Program details will be negotiated with a PI and their university/institution if their proposal is selected. Funding will be available only after we have agreed to the terms of and signed a sponsored research agreement.

Inquiries

Inquiries Related to the 2025 Sony Research Award Program

Please contact us via the Inquiry Form if you have a question regarding the 2025 Sony Research Award Program. The Research Award Program Administration Office is the only resource that can officially answer your question(s).

Inquiries Outside of the 2025 Sony Research Award Program

Please visit www.sony.com and contact the appropriate channel listed at the bottom of the web page if you have questions outside of the Sony Research Award Program (such as business proposals, joint venture proposals, research proposals that are not related to any of Focused Research Themes or keywords for the Faculty Innovation Award), or any other inquiry or proposal.

• *Awards will be paid in USD or respective currency depending on the country of the participating university/institution.
• **For European institutions, the contracting party will be the Sony entity/ branch of the country of the participating university/institution.

STORIES

Collaborative research with universities through the Sony Research Award Program continues to drive meaningful progress. The articles below highlight representative projects that are moving the field forward at various stages of development.