Professor Akella will describe his research, which is based on extensive interactions with Silicon Valley and firms such as AOL, SAP, Cisco, IBM.
He will focus particularly on Hierarchical Bayesian models, including Bayesian Kalman filtering, in addressing the following problems:
1. Online advertising and attribution modeling, which is assigning credit for user commercial actions to ad exposures. He will describe both aggregate methods and disaggregate methods that address incorrect A/B testing approaches. He will also describe Big data and sparsity issues.
2. Energy analytics for building energy optimization. He will indicate similarities in approaches based on Bayesian Kalman filtering.
3. Topic modeling, information extraction and retrieval in service centers.
The key issue is Big Data resulting in cognitive overload. We exploit the Generalized Dirichlet and other models that better capture causal models for text, and lead to significantly enhanced likelihood results with orders of magnitude speedup.
—————————–
Ram Akella is currently Professor and Director of the Center for Large-scale Live Analytics and Smart Services (CLASS), which includes SMART (Social Media Analytics Research Transformation), and the Center for Knowledge, Information Systems and Technology Management (KISMT) at the University of California at Silicon Valley Center/Santa Cruz. Prof. Akella started his academic career as a post-doctoral fellow at Harvard and then joined MIT (EECS/LIDS and LFM) as a Postdoctoral Associate. He then joined Carnegie Mellon University in 1985 as an Associate Professor in the Tepper Business School (GSIA) and the School of Computer Science (CS/RI), before working at other institutions including MIT, Berkeley, Stanford, and establishing an ORU and TIM(ISTM) at UCSC/SVC as Founding Director/Chair.
——————-
All talks may be viewed on our
The schedule for the semester can be found on the
Webviewing at UC Davis: 1003 Kemper Hall
Webviewing at UC Merced: SE1 100
Webviewing at UC Santa Cruz: SOE E2 Building, Room 595B