Publications

Abstract In the real world, many relationships between events are uncertain and probabilistic. Uncertainty is also likely to be a more common feature of daily experience for youth because they have less experience to draw from than adults. Some studies suggests probabilistic learning may be inefficient in youth compared to adults [1], while others suggest it may be more efficient in youth that are in mid adolescence [2, 3]. Here we used a probabilistic reinforcement learning task to test how youth age 8-17 (N = 187) and adults age 18-30 (N = 110) learn about stable probabilistic contingencies. Performance increased with age through early-twenties, then stabilized. Using hierarchical Bayesian methods to fit computational reinforcement learning models, we show that all participants’ performance was better explained by models in which negative outcomes had minimal to no impact on learning. The performance increase over age was driven by 1) an increase in learning rate (i.e. decrease in integration time horizon); 2) a decrease in noisy/exploratory choices. In mid-adolescence age 13-15, salivary testosterone and learning rate were positively related. We discuss our findings in the context of other studies and hypotheses about adolescent brain development. Author summary Adolescence is a time of great uncertainty. It is also a critical time for brain development, learning, and decision making in social and educational domains. There are currently contradictory findings about learning in adolescence. We sought to better isolate how learning from stable probabilistic contingencies changes during adolescence with a task that previously showed interesting results in adolescents. We collected a relatively large sample size (297 participants) across a wide age range (8-30), to trace the adolescent developmental trajectory of learning under stable but uncertain conditions. We found that age in our sample was positively associated with higher learning rates and lower choice exploration. Within narrow age bins, we found that higher saliva testosterone levels were associated with higher learning rates in participants age 13-15 years. These findings can help us better isolate the trajectory of maturation of core learning and decision making processes during adolescence.

Uniform clusters of Pt have been deposited on the surface of capping-agent-free CeO 2 nanooctahedra and nanorods using electron beam (e-beam) evaporation. The coverage of the Pt nanocluster layer can be controlled by adjusting the e-beam evaporation time. The resulting e-beam evaporated Pt nanocluster layers on the CeO 2 surfaces have a clean surface and clean interface between Pt and CeO 2. Different growth behaviors of Pt on the two types of CeO 2 nanocrystals were observed, with epitaxial growth of Pt on CeO 2 nanooctahedra and random growth of Pt on CeO 2 nanorods. The structures of the Pt clusters on the two different types of CeO 2 nanocrystals have been studied and compared by using them as catalysts for model reactions. The results of hydrogenation reactions clearly showed the clean and similar chemical surface of the Pt clusters in both catalysts. The support-dependent activity of these catalysts was demonstrated by CO oxidation. The Pt/CeO 2 nanorods showed much higher activity compared with Pt/CeO 2 nanooctahedra because of the higher concentration of oxygen vacancies in the CeO 2 nanorods. The structure-dependent selectivity of dehydrogenation reactions indicates that the structures of the Pt on CeO 2 nanorods and nanooctahedra are different. Thes differences arise because the metal deposition behaviors are modulated by the strong metal-metal oxide interactions. © 2011 Tsinghua University Press and Springer-Verlag Berlin Heidelberg.

Artificial intelligence (AI) has emerged as a tool for discovering and optimizing novel battery materials. However, the adoption of AI in battery cathode representation and discovery is still limited due to the complexity of optimizing multiple performance properties and the scarcity of high-fidelity data. In this study, we present a machine-learning model (DRXNet) for battery informatics and demonstrate the application in the discovery and optimization of disordered rocksalt (DRX) cathode materials. We have compiled the electrochemistry data of DRX cathodes over the past five years, resulting in a dataset of more than 19,000 discharge voltage profiles on diverse chemistries spanning 14 different metal species. Learning from this extensive dataset, our DRXNet model can automatically capture critical features in the cycling curves of DRX cathodes under various conditions. Illustratively, the model gives rational predictions of the discharge capacity for diverse compositions in the Li--Mn--O--F chemical space as well as for high-entropy systems. As a universal model trained on diverse chemistries, our approach offers a data-driven solution to facilitate the rapid identification of novel cathode materials, accelerating the development of next-generation batteries for carbon neutralization.

Initialization, normalization, and skip connections are believed to be three indispensable techniques for training very deep convolutional neural networks and obtaining state-of-the-art performance. This paper shows that deep vanilla ConvNets without normalization nor skip connections can also be trained to achieve surprisingly good performance on standard image recognition benchmarks. This is achieved by enforcing the convolution kernels to be near isometric during initialization and training, as well as by using a variant of ReLU that is shifted towards being isometric. Further experiments show that if combined with skip connections, such near isometric networks can achieve performances on par with (for ImageNet) and better than (for COCO) the standard ResNet, even without normalization at all. Our code is available at https://github.com/HaozhiQi/ISONet.

<title>Abstract</title> Synthetic biology enables microbial hosts to produce complex molecules that are otherwise produced by organisms that are rare or difficult to cultivate, but the structures of these molecules are limited to those formed by chemical reactions catalyzed by natural enzymes. The integration of artificial metalloenzymes (ArMs) that catalyze unnatural reactions into metabolic networks could broaden the cache of molecules produced biosynthetically by microorganisms. We report an engineered microbial cell expressing a heterologous biosynthetic pathway, which contains both natural enzymes and ArMs, that produces an unnatural product with high diastereoselectivity. To create this hybrid biosynthetic organism, we engineered Escherichia coli (E. coli) with a heterologous terpene biosynthetic pathway and an ArM containing an iridium-porphyrin complex that was transported into the cell with a heterologous transport system. We improved the diastereoselectivity and product titer of the unnatural product by evolving the ArM and selecting the appropriate gene induction and cultivation conditions. This work shows that synthetic biology and synthetic chemistry can produce, together with natural and artificial enzymes in whole cells, molecules that were previously inaccessible to nature.

Spinels are an important class of compounds in energy-dense multivalent batteries with potential cathode (MgMn 2 O 4 ) and solid electrolyte (MgIn 2 S 4 ) applications. Both MgMn 2 O 4 and MgIn 2 S 4 are susceptible to inversion, which is the exchange of Mg and metal (Mn/In) sites in the spinel framework. Using first-principles calculations, we evaluate the impact of spinel Inversion, which can directly lead to several distinct local cation environments, on Mg 2+ mobility. While activation barriers for various cation decorations determine the active Mg 2+ diffusion channels on the atomic scale, the macroscopic migration of Mg 2+ is essential for (dis)charge of cathodes or ionic conduction in solid electrolytes. Subsequently, we perform Monte-Carlo simulations to estimate the minimum Mg concentration required to ensure Mg 2+ percolation through the spinel structure. Finally, we analyze the influence of inversion on the electrochemical properties of the MgMn 2 O 4 cathode by investigating the phase behavior, average voltages and extractable capacities at various degrees of inversion.

Ni-(oxy)­hydroxide-based materials are promising earth-abundant catalysts for electrochemical water oxidation in basic media. Recent findings demonstrate that incorporation of trace Fe impurities from commonly used KOH electrolytes significantly improves oxygen evolution reaction (OER) activity over NiOOH electrocatalysts. Because nearly all previous studies detailing structural differences between α-Ni­(OH)2/γ-NiOOH and β-Ni­(OH)2/β-NiOOH were completed in unpurified electrolytes, it is unclear whether these structural changes are unique to the aging phase transition in the Ni-(oxy)­hydroxide matrix or if they arise fully or in part from inadvertent Fe incorporation. Here, we report an investigation of the effects of Fe incorporation on structure–activity relationships in Ni-(oxy)­hydroxide. Electrochemical, in situ Raman, X-ray photoelectron spectroscopy, and electrochemical quartz crystal microbalance measurements were employed to investigate Ni­(OH)2 thin films aged in Fe-free and unpurified (reagent-grade) 1 M KOH (&lt;1 ppm Fe). We find that Ni films aged in unpurified electrolyte can incorporate ≥20% Fe after 5 weeks of aging, and the maximum catalyst activity is comparable to that reported for optimized Ni1–x Fe x OOH catalysts. Conversely, Fe-free Ni­(OH)2 films exhibit a substantially lower activity and higher Tafel slope for the OER. Films aged in Fe-free electrolyte are predominantly disordered β-Ni­(OH)2/β-NiOOH if maintained below 0.7 V vs Hg/HgO in 1 M KOH and will “overcharge” to form a mixture of γ- and β-NiOOH above this potential. Fe-containing Ni­(OH)2 films evidence a lesser extent of β-Ni­(OH)2 formation and instead exhibit NiOOH structural changes in accordance with the formation of a NiFe-layered double hydroxide phase. Furthermore, turnover frequency calculations indicate that Fe is the active site within this phase, and above ∼11% Fe content, a separate, Fe-rich phase forms. These findings are the first to demonstrate the in situ changes in the catalyst structure resulting from the incorporation of Fe electrolyte impurities within Ni-(oxy)­hydroxide, providing direct evidence that a Ni–Fe layered double (oxy)­hydroxide (LDH) phase is critical for high OER activity.

personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission. Acknowledgement We would like to thank Mark Allman for his invaluable contribution to this work. IP Options are not an option