Research Areas | Yahoo Research

Applied Science and Fundamental Research

At Yahoo Research, we address unique challenges which require customized solutions that match our data scale and requirements. While crafting applications impacting millions, we actively contribute to and drive advancements in the scientific community. Our interests span various domains, including:

Advertising Science

We combine our immense trove of data with tools from large-scale machine learning, statistical modeling, feedback control, and more to develop the high-performing and robust algorithms that drive impression allocation and ad serving. including machine learning, statistics, operations research, economics, mathematics, control theory, etc. Some representative areas of work are: click and conversion prediction, viewability and video completion prediction, incremental reach optimization, campaign budget spend pacing, bidding strategies, supply path optimization, app marketing optimization, and optimization and performance measurement on internet traffic with privacy-driven restricted user traceability.

Generative AI

Our interdisciplinary team works tirelessly to pioneer inventive solutions to bring Generative AI to our customers. Through research and development, we architect cutting-edge algorithms and systems, including query generation, synthetic data generation, metadata enrichment, and multi-modal detection.

Image & Video Understanding

We research and innovate in image and video understanding, including object recognition, augmented-reality, visual aesthetics, and creativity. Our work directly impacts many of our flagship products and platforms by powering core technologies such as video highlights, labeling and summarization, and search by visual content.

Information Retrieval

We handle a lot of text: from mail and news, to travel reviews, user comments, and the whole web. Finding information and answers in them requires machine learning, natural language understanding and large-scale data processing.

Knowledge Graphs

As a team of research scientists and engineers, we develop methods and systems to extract, organize and enrich information about entities. Particularly, we extract information about entities from various sources of structured and unstructured data, integrate this information into large normalized knowledge graphs that contain billions of facts about millions of interconnected entities, and utilize these resulting knowledge graphs to create or enhance products and services across Yahoo, particularly for Web Search and Commerce use cases.

Machine Learning

Machine Learning is core to our products. Our researchers and engineers develop and deploy large scale ML and deep learning algorithms on one of the largest grid computing platforms in the world. We process over 100 billion daily user activities and over 200 billion daily ad requests over our global media and advertising networks.

Metrics and User Engagement

Metrics and User Engagement researchers collaborate across product teams to define metrics for experimentation, user engagement, and product success. We use data-driven user understanding methods to enable optimization of products through experimentation, and guide product development by using large-scale data in deep, grounded analysis.

Natural Language Understanding

Natural Language Understanding (NLU) encompasses statistical natural language understanding, dialogue planning, and understanding across multiple turns. We leverage language models to power classification, extraction, summarization, and dialogue management

Security

We tackle a broad spectrum of security challenges, including cybersecurity, privacy concerns, ethical considerations, and the impact of emerging technologies on society. Our researchers develop robust encryption techniques, intrusion detection systems, and access controls to fortify digital infrastructures against cyberattacks and data breaches.

Time Series Analysis & Forecasting

Our research and engineering team develop sophisticated mathematical models and algorithms to analyze historical data, uncover trends, seasonality, and dependencies, and make accurate predictions about future trends or values. We uncover financial insights to power products on Yahoo! Finance and detect popular trends to facilitate Yahoo! News and Search.

User Modeling and Personalization

Personalization is the science of developing user models to anticipate the interaction each user will have with a product. We look for ways in which personalization can be used within real-time experiences on mobile devices with rich context. Key problems in this area include contextual profiling, context-based intent prediction, personalized recommendations, and real-time federated search.

Other

Yahoo Research also explores unconventional ideas, niche domains, and interdisciplinary intersections that push the boundaries of technological innovation. Some notable examples are (but not limited to) Data Sketches and Search Monetization & Econometrics.

Publications

Read the latest Publications by our Yahoo Research team.

Browse Publications