Computer vision techniques have trouble recognizing subcategories of objects (for example, a vehicle’s model type or a bird’s species). A new method developed by ICSI researchers improves automatic recognition of subcategories by first warping small areas of photos to account for differences in pose and angle, and then grouping the areas according to their similarities.
Spammers who posted almost half a million Twitter messages in order to silence debate over Russia’s election in December likely purchased fraudulent accounts in bulk and posted the tweets from botnets, groups of malware-infected computers under the command of a single person. According to Networking Group researchers, the campaign took advantage of an underground economy based on spam, a phenomenon that researchers are studying in an attempt to improve methods of eliminating spam.
Charles Fillmore is one of the founding fathers of linguistics as it is practiced today. A professor emeritus of UC Berkeley, Chuck is also the director of ICSI’s FrameNet Project, which is building a lexical database, usable by both machines and humans, that describes the relationships between words in order to extract meaning from texts. The work requires the intellectual flexibility and passion for language that have marked his sixty-year career.
ICSI researchers and their collaborators have developed a way to statistically compare two of the most widely used lexical resources for English, FrameNet and WordNet. With the Expected Jaccard Index, researchers can for the first time empirically quantify agreement between WordNet and FrameNet expert annotations of words and phrases pulled from real-world texts. Comparing the annotations helps researchers find problematic gaps in the data and align the resources so they can be used together.
Professor Richard Karp has led ICSI’s Algorithms Group for the majority of its existence since 1988. Dick has just been named the founding director of an institute that will host research on theoretical computer science and its outreach to other scientific fields. The Simons Institute for the Theory of Computing will be funded by a $60 million grant to UC Berkeley from the Simons Foundation.
ICSI researchers and their collaborators have developed an algorithm that estimates where random Internet videos were shot by training itself on both videos for which locations have already been identified and those for which they haven’t. It’s the first multimodal location estimation method to use both training and test data to improve its results and, in some case, is over 10 percent more accurate than existing methods. The improvement is particularly high when few training videos are available.
Many notable computer scientists work at ICSI, and in recognition of their contributions to computer science, we will profile some of our current scientists and visitors, as well as notable alumni, here on the blog. We begin with Roberto Pieraccini, ICSI’s new director as of January 15, 2012. A speech scientist best known for applying statistical methods to spoken language understanding and dialog, Roberto has 30 years of experience in research and development at corporate laboratories and technology companies in the U.S. and Italy.
Welcome to the blog of the International Computer Science Institute (ICSI). On this blog, we will be sharing news about ICSI, profiles of our current and former researchers, and results from our work in computer networking and network security, speech and language processing, bioinformatics, computer architecture, computer vision, and artificial intelligence. Please feel free to leave comments or to contact us directly at [email protected].
ICSI is a leading center for research in computer science and one of the few independent, nonprofit research institutes in the U.S. We have a unique focus on international collaboration, bringing together computer science experts from all over the world. For more about us, visit our Web site at www.icsi.berkeley.edu.
Computer vision techniques have trouble recognizing subcategories of objects (for example, a vehicle’s model type or a bird’s species). A new method developed by ICSI researchers improves automatic recognition of subcategories by first warping small areas of photos to account for differences in pose and angle, and then grouping the areas according to their similarities.
Spammers who posted almost half a million Twitter messages in order to silence debate over Russia’s election in December likely purchased fraudulent accounts in bulk and posted the tweets from botnets, groups of malware-infected computers under the command of a single person. According to Networking Group researchers, the campaign took advantage of an underground economy based on spam, a phenomenon that researchers are studying in an attempt to improve methods of eliminating spam.
Charles Fillmore is one of the founding fathers of linguistics as it is practiced today. A professor emeritus of UC Berkeley, Chuck is also the director of ICSI’s 
ICSI researchers and their collaborators have developed a way to statistically compare two of the most widely used lexical resources for English, FrameNet and WordNet. With the Expected Jaccard Index, researchers can for the first time empirically quantify agreement between WordNet and FrameNet expert annotations of words and phrases pulled from real-world texts. Comparing the annotations helps researchers find problematic gaps in the data and align the resources so they can be used together.
Professor Richard Karp has led ICSI’s Algorithms Group for the majority of its existence since 1988. Dick has just been named the founding director of an institute that will host research on theoretical computer science and its outreach to other scientific fields. The Simons Institute for the Theory of Computing will be funded by a $60 million grant to UC Berkeley from the Simons Foundation.
ICSI researchers and their collaborators have developed an algorithm that estimates where random Internet videos were shot by training itself on both videos for which locations have already been identified and those for which they haven’t. It’s the first multimodal location estimation method to use both training and test data to improve its results and, in some case, is over 10 percent more accurate than existing methods. The improvement is particularly high when few training videos are available.