Dictionary learning algorithms have been successfully used for both reconstructive and discriminative tasks, where an input signal is represented with a sparse linear combination of dictionary atoms. While these methods are mostly developed for single-modality scenarios, recent studies have demonstrated the advantages of feature-level fusion based on the joint sparse representation of the multimodal inputs. In this paper, we propose a multimodal task-driven dictionary learning algorithm under the joint sparsity constraint (prior) to enforce collaborations among multiple homogeneous/heterogeneous sources of information. In this task-driven formulation, the multimodal dictionaries are learned simultaneously with their corresponding classifiers. The resulting multimodal dictionaries can generate discriminative latent features (sparse codes) from the data that are optimized for a given task such as binary or multiclass classification. Moreover, we present an extension of the proposed formulation using a mixed joint and independent sparsity prior, which facilitates more flexible fusion of the modalities at feature level. The efficacy of the proposed algorithms for multimodal classification is illustrated on four different applications--multimodal face recognition, multi-view face recognition, multi-view action recognition, and multimodal biometric recognition. It is also shown that, compared with the counterpart reconstructive-based dictionary learning algorithms, the task-driven formulations are more computationally efficient in the sense that they can be equipped with more compact dictionaries and still achieve superior performance.

Solution pH is a powerful tool for regulating many kinds of chemical activity, but is generally treated as a static property defined by a pre-selected buffer. Introducing dynamic control of pH in space, time, and magnitude can enable richer and more efficient chemistries, but is not feasible with traditional methods of titration or buffer exchange. Recent reports have featured electrochemical strategies for modifying bulk pH in constrained volumes, but only demonstrate switching between two preset values and omit spatial control entirely. Here, we use a combination of solution-borne quinones and galvanostatic excitation to enable quantitative control of pH environments that are highly localized to an electrode surface. We demonstrate highly reproducible acidification and alkalinization with up to 0.1 pH s(-1) (±0.002 pH s(-1)) rate of change across the dynamic range of our pH sensor (pH 4.5 to 7.5) in buffered solutions. Using dynamic current control, we generate and sustain 3 distinct pH microenvironments simultaneously to within ±0.04 pH for 13 minutes in a single solution, and we leverage these microenvironments to demonstrate spatially-resolved, pH-driven control of enzymatic activity. In addition to straightforward applications of spatio-temporal pH control (e.g. efficiently studying pH-dependencies of chemical interactions), the technique opens completely new avenues for implementing complex systems through dynamic control of enzyme activation, protein binding affinity, chemical reactivity, chemical release, molecular self-assembly, and many more pH-controlled processes.