John Bennett

Learn More
Much research in handwriting recognition has focused on how to improve recognizers with constrained training set sizes. This paper presents the results of training a nearest-neighbor based online Japanese Kanji recognizer and a neural-network based online cursive English recognizer on a wide range of training set sizes, including sizes not generally(More)
We describe the role of mentors in an experimental course on human-computer interaction (HCI) taught in the Computer Science Department at Stanford University for the past two years. Students practice design within the course by collaborating in small groups on 12-week projects, in which they analyze a work environment, design and implement a prototype user(More)
Since the CHI community involves both researchers and practitioners, we often sQuggle with the issue of technology transfer. The CHI conference features many innovative research ideas and interesting product designs, but there have been disappointingly few cases in which products were based on research projects. Although many compauies have tried to address(More)
1 Abstract-This paper describes an efficient model for twisted-pair cables suitable for discrete time simulations. The model uses a pole-zero approximation for the cable losses, and a modified s-parameter approach to model the reflections due to dis-continuities and stubs. Use of this model can improve simulation times by a factor of ten compared to using(More)
  • Cu Scholar, John Giacomoni, John Bennett, Antonio Carzaniga, Manish Vachharajani, John Giacomoni +5 others
  • 2015
The high performance, low cost, and flexibility of commodity hardware systems make them appealing for network processing applications. However, the standard software architecture of such systems imposes significant limitations. At high rates (e.g., Gigabit Ethernet) and small frame sizes (64 byte) each frame must be processed in less than 672 ns. System(More)
— We have developed a novel control mechanism that deploys a large number of inexpensive robots as a distributed remote sensing array, called a Distributed Robotic Macrosen-sor (DRM). A simple virtual spring mesh abstraction is used to provide fully distributed control that is both flexible and fault-tolerant. We describe and evaluate several algorithms for(More)