Publications

Water-vapor fed electrolysis, a simplified single-phase electrolyzer using a proton-exchange membrane electrode assembly, achieved >100 mA cm −2 performance at <1.7 V, the best for water-vapor electrolysis to date, and was tested under various operating conditions (temperature and inlet relative humidity (RH)). To further probe the limitations of the electrolyzer, a mathematical model was used to identify the overpotentials, local water activity, water content values, and temperature within the cell at these various conditions. The major limitations within the water-vapor electrolyzer are caused by a decreased water content within the membrane phase, indicated by increased Ohmic and mass transport losses seen in applied voltage breakdowns. Further investigations show the water content ( λ , mole of water/mole of sulfonic acid) can decrease from 13 at low current densities down to 6 at high current densities. Increasing the temperature or decreasing RH exacerbates this dry-out effect. Using our mathematical model, we show how these mass transport limitations can be alleviated by considering the role of water as both a reactant and a hydrating agent. We show that low cathode RH can be tolerated as long as the anode RH remains high, showing equivalent performance as symmetric RH feeds.

We report the development of a modular, one-pot, sequential chemoenzymatic system for the formal enantioselective construction of the C–C bond in 2-aryl 1,4-dicarbonyl compounds. This sequence comprises a rhodium-catalyzed diazocoupling that provides >9:1 selectivity for heterocoupling of two diazoesters and a reduction mediated by an ene-reductase (ER), which occurs in up to 99% enantiomeric excess (ee). The high yield and enantioselectivity of this system result from the preferential generation of an (E)-alkene from the diazo coupling reaction and selective reduction of the (E)-alkene in a mixture of (E) and (Z) isomers by the ER. Screening of a panel of ERs revealed that OPR1 from Lycopersicum esculentum catalyzes the reduction of bulky tert-butyl or benzyl esters to afford chiral diesters that are poised for orthogonal reactions at the two distinct ester units of the product. Overall, this work demonstrates the benefit of combining organometallic and enzymatic catalysis to create unusual overall transformations that do not require the isolation and purification of intermediates.

The goal of this work is to detect a human figure image and localize his joints and limbs along with their associated pixel masks. In this work we attempt to tackle this problem in a general setting. The dataset we use is a collection of sports news photographs of baseball players, varying dramatically in pose and clothing. The approach that we take is to use segmentation to guide our recognition algorithm to salient bits of the image. We use this segmentation approach to build limb and torso detectors, the outputs of which are assembled into human figures. We present quantitative results on torso localization, in addition to shortlisted full body configurations.

An entry from the Cambridge Structural Database, the world’s repository for small molecule crystal structures. The entry contains experimental data from a crystal diffraction study. The deposited dataset for this entry is freely available from the CCDC and typically includes 3D coordinates, cell parameters, space group, experimental conditions and quality measures.

Machine learning applications are increasingly deployed not only to serve predictions using static models, but also as tightly-integrated components of feedback loops involving dynamic, real-time decision making. These applications pose a new set of requirements, none of which are difficult to achieve in isolation, but the combination of which creates a challenge for existing distributed execution frameworks: computation with millisecond latency at high throughput, adaptive construction of arbitrary task graphs, and execution of heterogeneous kernels over diverse sets of resources. We assert that a new distributed execution framework is needed for such ML applications and propose a candidate approach with a proof-of-concept architecture that achieves a 63x performance improvement over a state-of-the-art execution framework for a representative application.

A triphenylphosphinegold(I)-catalyzed cyclopropanation of olefins using propargyl esters as gold(I)-carbene precursors is reported. This reaction provided the basis for the use of a DTBM-SEGPHOS gold(I) complex as a catalyst in the enantioselective (up to 94% ee) preparation of vinyl cyclopropanes with high cis-selectivity.

We have demonstrated that cascaded fiber Bragg gratings can provide delays of propagating pulses with minimal pulse reshaping. The grating pair used exhibited an overlap transmission region centered at 1551.05 nm, where both gratings contribute to the group delay and the group velocity dispersion (GVD) was canceled. Using wavelength tunable pulses spectrally sliced from a mode-locked fiber laser, the measurement was performed in the time domain with single picosecond resolution. Both gratings were 3 mm long and a maximum group delay of 15 ps was measured for the cascaded sequence. This compound grating configuration can be implemented as encoders and decoders in spread spectrum CDMA systems.