Personalization is a long-standing problem in data mining and machine learning. Companies make personalized product recommendations to millions of users every second. In addition to the recommendation problem, with the emerging of personal devices, many conventional problems, e.g., recognition, need to be personalized as well. Moreover, as the number of users grows huge, solving personalization becomes quite challenging. In this paper, we formalize the generic personalization problem as an optimization problem. We propose several ADMM algorithms to solve this problem in a distributed way including a new Asynchronous ADMM that removes all synchronous barriers to maximize the training throughput. We provide a mathematical analysis to show that the proposed Asynchronous ADMM algorithm holds a linear convergence rate which is the best to our knowledge. The distributed personalization allows training to be performed in either a cluster or even on a user's device. This can improve the privacy protection as no personal data is uploaded, while personal models can still be shared with each other. We apply this approach to two industry problems, \emph{Facial Expression Recognition} and \emph{Job Recommendation}. Experiments demonstrate more than 30\% relative error reduction on both problems. Asynchronous ADMM allows faster training for problems with millions of users since it eliminates all network I/O waiting time to maximize the cluster CPU throughput. Experiments demonstrate 4 times faster than original synchronous ADMM algorithm.

Data annotation bias is found in many situations. Often it can be ignored as just another component of the noise floor. However, it is especially prevalent in crowdsourcing tasks and must be actively managed. Annotation bias on single data items has been studied with regard to data difficulty, annotator bias, etc., while annotation bias on batches of multiple data items simultaneously presented to annotators has not been studied. In this paper, we verify the existence of "in-batch annotation bias" between data items in the same batch. We propose a factor graph based batch annotation model to quantitatively capture the in-batch annotation bias, and measure the bias during a crowdsourcing annotation process of inappropriate comments in LinkedIn. We discover that annotators tend to make polarized annotations for the entire batch of data items in our task. We further leverage the batch annotation model to propose a novel batch active learning algorithm. We test the algorithm on a real crowdsourcing platform and find that it outperforms in-batch bias naïve algorithms.

DNA sequencing by hybridization, proposed about a decade ago as an alternative to standard electrophoresis-based sequencing techniques, is relevant not only to sequencing per se but also to diagnostics and therapeutics. The inherent structural inadequacy of traditional probe patterns and well-known hybridization shortcomings had for some time deflated the interest in this approach. Renewed interest for this topic has been generated by our recent discovery of a novel probing scheme whose performance for the first time asymptotically meets the information theory bound. After settling the question of asymptotic performance, due to the sizable volume of potential applications, the research focus has naturally shifted to issues of algorithmic fine tunings aimed at improving the performance "constants". In this paper, we introduce as a figure of merit the nucleotide-per-feature ratio (for a given confidence level), i.e., the length of a reliabily reconstructible (random) sequence divided by the number of probes placed on the sequencing array, regardless of the fabrication technology, and we explore the capabilities offered by the joint use of multiple different arrays in sequencing the same DNA sequence. In particular we consider a probing scheme based on the use of the spectra pertaining to a given probing pattern and to its reversal (referred to here as tandem spectra), and we show analytically and experimentally a performance improvement per unit of microarray area of about 4/3 (in a representative instance) over the conventional single-spectrum approach. The proposed tandem-spectrum reconstruction, in conjunction with a "voting provision" discussed elsewhere (whereby one of two competing alternatives is chosen if the corresponding probes are less numerous in the current prefix of the reconstructive sequence), is the best known technique for sequencing by hybridization.