Google researchers write a lot of papers on topics including computing, data, statistics and more, and share them often with the world’s scientific community to drive innovation and new ideas.
As part of that ongoing effort, Google is now sharing some of the most influential research papers produced by Google researchers in 2013 to expand discussions and observations on a raft of topics being studied around the world. The collection of research papers was announced by Corinna Cortes and Alfred Spector of the Google Research team in a June 30 post on the Google Research Blog.
“Googlers across the company actively engage with the scientific community by publishing technical papers, contributing open-source packages, working on standards, introducing new APIs and tools, giving talks and presentations, participating in ongoing technical debates, and much more,” wrote Cortes and Spector. “Our publications offer technical and algorithmic advances, feature aspects we learn as we develop novel products and services, and shed light on some of the technical challenges we face at Google.”
So what are some of the most influential Google research papers from 2013?
There’s a report on robotics titled “Cloud-based robot grasping with the google object recognition engine,” which looks at how robotics of the future could work with cloud-based controls rather than with on-board controls. The report looks at how this could be possible using wireless networking and rapidly expanding cloud computing resources.
Also tackled is the topic of distributed systems in the paper, “Photon: Fault-Tolerant and Scalable Joining of Continuous Data Streams,” which looks at Photon, a geographically distributed system for joining multiple continuously flowing streams of data in real-time with high scalability and low latency, according to the post. “The streams may be unordered or delayed. Photon fully tolerates infrastructure degradation and data center-level outages without any manual intervention while joining every event exactly once. Photon is currently deployed in production, processing millions of events per minute at peak with an average end-to-end latency of less than 10 seconds.”
Touch-screen research involving human-computer interaction is the topic of a paper, “FFitts Law: Modeling Finger Touch with Fitts’ Law, which looks to “more reliably model touch-screen target acquisition with finger touch,” as researchers continue to seek ways of making such interactions more natural for users.
Machine learning is a continuing topic, as seen in papers including “Ad Click Prediction: a View from the Trenches,” which provides a case study for ad click prediction; and in the paper “Efficient Estimation of Word Representations in Vector Space, which looks at a “simple and speedy method for training vector representations of words,” according to the post.
“The resulting vectors naturally capture the semantics and syntax of word use, such that simple analogies can be solved with vector arithmetic. For example, the vector difference between ‘man’ and ‘woman’ is approximately equal to the difference between ‘king’ and ‘queen’, and vector displacements between any given country’s name and its capital are aligned,” the post read.
Natural language processing is the topic of the paper, “Token and Type Constraints for Cross-Lingual Part-of-Speech Tagging,” which looks at new techniques in this field. “Knowing the parts of speech (verb, noun, etc.) of words is important for many natural language processing applications, such as information extraction and machine translation,” the post states.
Google Shares Research Findings With Scientific World
“Constructing part-of-speech taggers typically requires large amounts of manually annotated data, which is missing in many languages and domains. In this paper, we introduce a method that instead relies on a combination of incomplete annotations projected from English with incomplete crowd-sourced dictionaries in each target language. The result is a 25 percent error reduction compared to the previous state of the art,” the post continues.
Computer networks are the topic of the paper, “B4: Experience with a Globally Deployed Software Defined WAN,” which “presents the motivation, design, and evaluation of B4, a software defined WAN” for data center to data center connectivity, according to the post. The authors of the report present their approach to separating the network’s control plane from the data plane to enable rapid deployment of new network control services.
The security concerns surrounding data localization and the cloud are the topic in the paper, “When the Cloud Goes Local: The Global Problem With Data Localization,” according to the post. “Ongoing efforts to legally define cloud computing and regulate separate parts of the Internet are unlikely to address underlying concerns about data security and privacy,” wrote the paper’s authors. “Data localization initiatives, led primarily by European countries, could actually bring the cloud to the ground and make the Internet less secure.”
Statistics research was also covered in Google’s top research for 2013. In the report, “Pay by the Bit: An Information-Theoretic Metric for Collective Human Judgment,” a researcher looked at the topic of quality control in crowdsourcing. The report found that information theory provides a natural and elegant metric for the value of contributors’ work, in the form of the mutual information between their judgments and the questions’ answers, each treated as random variables.
Google and its staff are often working on research in many topics in computing.
In May 2014, Google launched a new Quantum Artificial Intelligence Lab to find ways to make computers much smarter so they can help solve some of the world’s most challenging problems, from diseases to environmental threats.
In February, it announced its first-ever Google App Engine Research Awards to seven projects that will use the App Engine platform’s abilities to work with large data sets for academic and scientific research. The new program, which was announced in the spring of 2012, brought in many proposals for a wide variety of scientific research, including in subject areas such as mathematics, computer vision, bioinformatics, climate and computer science.
Google created the fledgling App Engine Research Awards program to bolster its support of academic research, while providing academic researchers with access to Google’s infrastructure so they can explore innovative ideas in their fields, according to Google. The App Engine platform is designed for managing heavy data loads and running large-scale applications.