In recent years online learning (sequential prediction) has received much attention as it often produces fast and simple learning algorithms that enjoy robustness to changing or even adversarial data sources. However, despite the extensive existing literature on online learning, our theoretical understanding of the framework has been rather lacking. Most existing analyses have been case by case, and there is a lack of a general theory and methodology for designing online learning algorithms for the problem at hand.
Modern wireless technology enables the vision of future large scale systems such as the SmartGrid, a network of ubiquitous and heterogeneous devices wirelessly connected to the Internet, and wireless health monitoring and health modifying sensor networks over communities and not just individuals. All of these applications necessitate methods that simultaneously consider scale, communication, sensing and control. In this talk, key elements of realizing this vision are examined.
Power dissipation and off-chip bandwidth restrictions are critical challenges that limit microprocessor performance. Ternary content addressable memories (TCAM) hold the potential to address both problems in the context of a wide range of data-intensive workloads that benefit from associative search capability. Power dissipation is reduced by eliminating instruction processing and data movement overheads present in a purely RAM based system. Bandwidth demand is lowered by processing data directly on the TCAM chip, thereby decreasing off-chip traffic.
Host cells undergo rapid and drastic transcriptional changes to mount a
response to pathogen infection. The response is triggered when upstream host
proteins detect the pathogen and propagate signals via cascades of
protein-protein interactions to transcription factors, which selectively
activate or inhibit genes. Although these upstream initiators and the
resulting transcriptional effects can be characterized experimentally, the
intermediate proteins driving signaling and transcription cannot be directly
Chaperones are special proteins that aid the folding, unfolding, assembly
and disassembly of other proteins. Chaperones rely on a large and diverse
set of co-chaperones that regulate their specificity and function. How the
se co-chaperones regulate protein folding and whether they have
chaperone-independent biological functions is largely unknown. In this talk,
I will first present novel experimental and statistical approaches to study
the chaperone/co-chaperone/client inter action network in a systematic way.
Affordable mobile devices and ubiquitous wireless connectivity have placed digital communication, computation, and sensing at the center of nearly all human activity. However, as the importance and popularity of mobile services like citizen journalism and mobile social networking have grown, so too has the need to reason about the trustworthiness of the data on which these services depend.
These days, there is a lot of hype about big data. But such data is useless
unless we can interpret it. Since such data is often noisy and ambiguous,
especially in the "long tail" where many of the examples occur, we will
inevitably be somewhat uncertain about underlying patterns and/or future
predictions. This motivates analyzing the data using structured
probabilistic models, which can properly represent uncertainty, and exploit
prior knowledge when available.