Browsing by Author "Chen L"

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Analytical Investigations into Anomalous Diffusion Driven by Stress Redistribution Events: Consequences of Lévy Flights
    (MDPI (Basel, Switzerland), 2022-09-06) Cleland JD; Williams MAK; Lopes AM; Alfi A; Chen L; David SA
    This research is concerned with developing a generalised diffusion equation capable of describing diffusion processes driven by underlying stress-redistributing type events. The work utilises the development of an appropriate continuous time random walk framework as a foundation to consider a new generalised diffusion equation. While previous work has explored the resulting generalised diffusion equation for jump-timings motivated by stick-slip physics, here non-Gaussian probability distributions of the jump displacements are also considered, specifically Lévy flights. This work illuminates several features of the analytic solution to such a generalised diffusion equation using several known properties of the Fox H function. Specifically demonstrated are the temporal behaviour of the resulting position probability density function, and its normalisation. The reduction of the proposed form to expected known solutions upon the insertion of simplifying parameter values, as well as a demonstration of asymptotic behaviours, is undertaken to add confidence to the validity of this equation. This work describes the analytical solution of such a generalised diffusion equation for the first time, and additionally demonstrates the capacity of the Fox H function and its properties in solving and studying generalised Fokker–Planck equations.
  • Thumbnail Image
    Item
    Anisotropic span embeddings and the negative impact of higher-order inference for coreference resolution: An empirical analysis
    (Cambridge University Press, 2024-01-25) Hou F; Wang R; Ng S-K; Zhu F; Witbrock M; Cahan SF; Chen L; Jia X
    Coreference resolution is the task of identifying and clustering mentions that refer to the same entity in a document. Based on state-of-the-art deep learning approaches, end-to-end coreference resolution considers all spans as candidate mentions and tackles mention detection and coreference resolution simultaneously. Recently, researchers have attempted to incorporate document-level context using higher-order inference (HOI) to improve end-to-end coreference resolution. However, HOI methods have been shown to have marginal or even negative impact on coreference resolution. In this paper, we reveal the reasons for the negative impact of HOI coreference resolution. Contextualized representations (e.g., those produced by BERT) for building span embeddings have been shown to be highly anisotropic. We show that HOI actually increases and thus worsens the anisotropy of span embeddings and makes it difficult to distinguish between related but distinct entities (e.g., pilots and flight attendants). Instead of using HOI, we propose two methods, Less-Anisotropic Internal Representations (LAIR) and Data Augmentation with Document Synthesis and Mention Swap (DSMS), to learn less-anisotropic span embeddings for coreference resolution. LAIR uses a linear aggregation of the first layer and the topmost layer of contextualized embeddings. DSMS generates more diversified examples of related but distinct entities by synthesizing documents and by mention swapping. Our experiments show that less-anisotropic span embeddings improve the performance significantly (+2.8 F1 gain on the OntoNotes benchmark) reaching new state-of-the-art performance on the GAP dataset.
  • Thumbnail Image
    Item
    How Reproducible are Surface Areas Calculated from the BET Equation?
    (Wiley-VCH GmbH, 2022-05-23) Osterrieth JWM; Rampersad J; Madden D; Rampal N; Skoric L; Connolly B; Allendorf MD; Stavila V; Snider JL; Ameloot R; Marreiros J; Ania C; Azevedo D; Vilarrasa-Garcia E; Santos BF; Bu X-H; Chang Z; Bunzen H; Champness NR; Griffin SL; Chen B; Lin R-B; Coasne B; Cohen S; Moreton JC; Colón YJ; Chen L; Clowes R; Coudert F-X; Cui Y; Hou B; D'Alessandro DM; Doheny PW; Dincă M; Sun C; Doonan C; Huxley MT; Evans JD; Falcaro P; Ricco R; Farha O; Idrees KB; Islamoglu T; Feng P; Yang H; Forgan RS; Bara D; Furukawa S; Sanchez E; Gascon J; Telalović S; Ghosh SK; Mukherjee S; Hill MR; Sadiq MM; Horcajada P; Salcedo-Abraira P; Kaneko K; Kukobat R; Kenvin J; Keskin S; Kitagawa S; Otake K-I; Lively RP; DeWitt SJA; Llewellyn P; Lotsch BV; Emmerling ST; Pütz AM; Martí-Gastaldo C; Padial NM; García-Martínez J; Linares N; Maspoch D; Suárez Del Pino JA; Moghadam P; Oktavian R; Morris RE; Wheatley PS; Navarro J; Petit C; Danaci D; Rosseinsky MJ; Katsoulidis AP; Schröder M; Han X; Yang S; Serre C; Mouchaham G; Sholl DS; Thyagarajan R; Siderius D; Snurr RQ; Goncalves RB; Telfer S; Lee SJ; Ting VP; Rowlandson JL; Uemura T; Iiyuka T; van der Veen MA; Rega D; Van Speybroeck V; Rogge SMJ; Lamaire A; Walton KS; Bingel LW; Wuttke S; Andreo J; Yaghi O; Zhang B; Yavuz CT; Nguyen TS; Zamora F; Montoro C; Zhou H; Kirchon A; Fairen-Jimenez D
    Porosity and surface area analysis play a prominent role in modern materials science. At the heart of this sits the Brunauer-Emmett-Teller (BET) theory, which has been a remarkably successful contribution to the field of materials science. The BET method was developed in the 1930s for open surfaces but is now the most widely used metric for the estimation of surface areas of micro- and mesoporous materials. Despite its widespread use, the calculation of BET surface areas causes a spread in reported areas, resulting in reproducibility problems in both academia and industry. To prove this, for this analysis, 18 already-measured raw adsorption isotherms were provided to sixty-one labs, who were asked to calculate the corresponding BET areas. This round-robin exercise resulted in a wide range of values. Here, the reproducibility of BET area determination from identical isotherms is demonstrated to be a largely ignored issue, raising critical concerns over the reliability of reported BET areas. To solve this major issue, a new computational approach to accurately and systematically determine the BET area of nanoporous materials is developed. The software, called "BET surface identification" (BETSI), expands on the well-known Rouquerol criteria and makes an unambiguous BET area assignment possible.
  • Thumbnail Image
    Item
    The LONELY GUY gene family: from mosses to wheat, the key to the formation of active cytokinins in plants
    (John Wiley and Sons Ltd on behalf of the Society for Experimental Biology and The Association of Applied Biologists, 2022-04-07) Chen L; Jameson GB; Guo Y; Song J; Jameson PE
    LONELY GUY (LOG) was first identified in a screen of rice mutants with defects in meristem maintenance. In plants, LOG codes for cytokinin riboside 5'-monophosphate phosphoribohydrolase, which converts inactive cytokinin nucleotides directly to the active free bases. Many enzymes with the PGGxGTxxE motif have been misannotated as lysine decarboxylases; conversely not all enzymes containing this motif are cytokinin-specific LOGs. As LOG mutants clearly impact yield in rice, we investigated the LOG gene family in bread wheat. By interrogating the wheat (Triticum aestivum) genome database, we show that wheat has multiple LOGs. The close alignment of TaLOG1, TaLOG2 and TaLOG6 with the X-ray structures of two functional Arabidopsis thaliana LOGs allows us to infer that the wheat LOGs 1-11 are functional LOGs. Using RNA-seq data sets, we assessed TaLOG expression across 70 tissue types, their responses to various stressors, the pattern of cis-regulatory elements (CREs) and intron/exon patterns. TaLOG gene family members are expressed variously across tissue types. When the TaLOG CREs are compared with those of the cytokinin dehydrogenases (CKX) and glucosyltransferases (CGT), there is close alignment of CREs between TaLOGs and TaCKXs reflecting the key role of CKX in maintaining cytokinin homeostasis. However, we suggest that the main homeostatic mechanism controlling cytokinin levels in response to biotic and abiotic challenge resides in the CGTs, rather than LOG or CKX. However, LOG transgenics and identified mutants in rice variously impact yield, providing interesting avenues for investigation in wheat.