Publications

The design process of user interfaces (UIs) often begins with articulating high-level design goals. Translating these high-level design goals into concrete design mock-ups, however, requires extensive effort and UI design expertise. To facilitate this process for app designers and developers, we introduce three deep-learning techniques to create low-fidelity UI mock-ups from a natural language phrase that describes the high-level design goal (e.g. "pop up displaying an image and other options"). In particular, we contribute two retrieval-based methods and one generative method, as well as pre-processing and post-processing techniques to ensure the quality of the created UI mock-ups. We quantitatively and qualitatively compare and contrast each method's ability in suggesting coherent, diverse and relevant UI design mock-ups. We further evaluate these methods with 15 professional UI designers and practitioners to understand each method's advantages and disadvantages. The designers responded positively to the potential of these methods for assisting the design process.

Imine-linked ILCOF-1 based on 1,4-phenylenediamine and 1,3,6,8-tetrakis(4-formylphenyl)pyrene was converted through consecutive linker substitution and oxidative cyclization to two isostructural covalent organic frameworks (COFs), having thiazole and oxazole linkages. The completeness of the conversion was assessed by infrared and solid-state NMR spectroscopies, and the crystallinity of the COFs was confirmed by powder X-ray diffraction. Furthermore, the azole-linked COFs remain porous, as shown by nitrogen sorption experiments. The materials derived in this way demonstrate increased chemical stability, relative to the imine-linked starting material. This constitutes a facile method for accessing COFs and linkages that are otherwise difficult to crystallize due to their inherently limited microscopic reversibility.

Intelligently extracting and linking complex scientific information from unstructured text is a challenging endeavor particularly for those inexperienced with natural language processing. Here, we present a simple sequence-to-sequence approach to joint named entity recognition and relation extraction for complex hierarchical information in scientific text. The approach leverages a pre-trained large language model (LLM), GPT-3, that is fine-tuned on approximately 500 pairs of prompts (inputs) and completions (outputs). Information is extracted either from single sentences or across sentences in abstracts/passages, and the output can be returned as simple English sentences or a more structured format, such as a list of JSON objects. We demonstrate that LLMs trained in this way are capable of accurately extracting useful records of complex scientific knowledge for three representative tasks in materials chemistry: linking dopants with their host materials, cataloging metal-organic frameworks, and general chemistry/phase/morphology/application information extraction. This approach represents a simple, accessible, and highly-flexible route to obtaining large databases of structured knowledge extracted from unstructured text. An online demo is available at http://www.matscholar.com/info-extraction.

We construct a data set of metal-organic framework (MOF) linkers and employ a fine-tuned GPT assistant to propose MOF linker designs by mutating and modifying the existing linker structures. This strategy allows the GPT model to learn the intricate language of chemistry in molecular representations, thereby achieving an enhanced accuracy in generating linker structures compared with its base models. Aiming to highlight the significance of linker design strategies in advancing the discovery of water-harvesting MOFs, we conducted a systematic MOF variant expansion upon state-of-the-art MOF-303 utilizing a multidimensional approach that integrates linker extension with multivariate tuning strategies. We synthesized a series of isoreticular aluminum MOFs, termed Long-Arm MOFs (LAMOF-1 to LAMOF-10), featuring linkers that bear various combinations of heteroatoms in their five-membered ring moiety, replacing pyrazole with either thiophene, furan, or thiazole rings or a combination of two. Beyond their consistent and robust architecture, as demonstrated by permanent porosity and thermal stability, the LAMOF series offers a generalizable synthesis strategy. Importantly, these 10 LAMOFs establish new benchmarks for water uptake (up to 0.64 g g^-1) and operational humidity ranges (between 13 and 53%), thereby expanding the diversity of water-harvesting MOFs.

4.1 Research Agenda: NEBULA Data Plane (NDP)................................................................................ 5 4.2 What is NDP?.................................................................................................................................... 6 Figure 1: NDP packet format....................................................................................................................... 7 4.3 What properties does NDP uphold?............................................................................................... 7

We recently described a new method to determine total body creatine pool size and skeletal muscle mass based on dilution of an oral dose of D 3 ‐creatine in urinary creatinine. Enrichment of D 3 ‐ creatinine was measured by isotope ratio mass spectrometry. Here we adapt the method to LC‐MS/MS, and investigate the potential for repeated application over time. Rats were given an oral tracer dose of D 3 ‐creatine (1.0 mg/kg body weight) at 10 weeks of age. Urine D 3 ‐creatinine enrichment was determined 72 hr after tracer administration and creatine pool size was calculated. Lean body mass (LBM) measured by quantitative magnetic resonance correlated with creatine pool size ( r = 0.92, P = 0.0002). The rate constant for decline in urinary D 3 ‐creatinine enrichment was slow and linear (2.73 ± 0.06 %/day). Subtracting background urinary D 3 ‐creatinine enrichment from the enrichment following a second dose of D 3 ‐creatine at 17 weeks permitted repeat calculations of creatine pool size. As at 10 weeks, 17 week LBM correlated with creatine pool size ( r = 0.98, P < 0.0001). In addition, the change in creatine pool size was correlated with the change in LBM during the 7 weeks of somatic growth between measurements ( r = 0.96, P < 0.0001). Thus, the LC‐MS/MS‐based D 3 ‐creatine dilution method can be applied repeatedly to measure skeletal muscle mass change in longitudinal studies. These studies were funded by GlaxoSmithKline.

Playing Fields presents the profound reflections of a group of international scholars on how games, sports, and motor practices interact with global-local processes, inequality, gender relations, identity, representation, performance, and emotion through varied modes of analysis, approaches, and styles.

Aims. The association of Type Ic supernovae (SNe) with long-duration gamma-ray bursts (GRBs) is well established. We endeavor, through\naccurate ground-based observational campaigns, to characterize these SNe at increasingly high redshifts.\nMethods. We obtained a series of optical photometric and spectroscopic observations of the Type Ic SN 2012bz associated with the Swift long-\nduration GRB 120422A (redshift z = 0.283) using the 3.6-m TNG and the 8.2-m VLT telescopes during the time interval between 4 and 36 days\nafter the burst.\nResults. The peak times of the light curves of SN 2012bz in various optical filters differ, with the B-band and i'-band light curves reaching\nmaximum at 9 ± 4 and 23 ± 3 rest-frame days, respectively. The bolometric light curve has been derived from individual bands photometric\nmeasurements, but no correction for the unknown contribution in the near-infrared (probably around 10­-15%) has been applied. Therefore, the\npresent light curve should be considered as a lower limit to the actual UV-optical-IR bolometric light curve. This pseudo-bolometric curve reaches\nits maximum (Mbol = -18.56 ± 0.06) at 13 ± 1 rest-frame days; it is similar in shape and luminosity to the bolometric light curves of the SNe\nassociated with z < 0.2 GRBs and more luminous than those of SNe associated with X-ray flashes (XRFs). A comparison with the model generated\nfor the bolometric light curve of SN 2003dh suggests that SN 2012bz produced only about 15% less 56 Ni than SN 2003dh, about 0.35 M . Similarly\nthe VLT spectra of SN 2012bz, after correction for Galactic extinction and for the contribution of the host galaxy, suggest comparable explosion\nparameters with those observed in SN 2003dh (EK 3.5 × 1052 erg, Mej 7 M ) and a similar progenitor mass (25­40 M ). GRB 120422A is\nconsistent with the Epeak - Eiso and the EX,iso - E,iso - Epeak relations. GRB 120422A / SN 2012bz shows the GRB-SN connection at the highest\nredshift so far accurately monitored both photometrically and spectroscopically.

MOFs with Guest Rooms Metal organic framework (MOF) solids typically boast large pores in their crystal lattices that can accommodate molecular guests and are useful as compact media for gas storage. However, guest binding has generally relied upon nonspecific interactions with the framework components. Li et al. (p. 855 ) have now prepared MOFs that incorporate macrocyclic ethers into the structural ligands comprising the framework walls within which certain cationic guests can bind quantitatively and site-specifically, akin to the molecular docking of drug molecules within protein receptors.

We introduce a novel sequential modeling approach which enables learning a Large Vision Model (LVM) without making use of any linguistic data. To do this, we define a common format, "visual sentences", in which we can represent raw images and videos as well as annotated data sources such as semantic segmentations and depth reconstructions without needing any meta-knowledge beyond the pixels. Once this wide variety of visual data (comprising 420 billion tokens) is represented as sequences, the model can be trained to minimize a cross-entropy loss for next token prediction. By training across various scales of model architecture and data diversity, we provide empirical evidence that our models scale effectively. Many different vision tasks can be solved by designing suitable visual prompts at test time.