Fine-grained Units Research Articles

Multi-stage magma emplacement in the western end of the Jinchuan Ni[sbnd]Cu sulfide deposit, NW China, and mineral exploration strategy based on exploration data

The western intrusion of the world-class Jinchuan NiCu sulfide deposit consists of fine-grained and coarse-grained binary lithofacies units. However, recent exploration has revealed two layers of net-textured sulfides in the fine-grained unit at the western end of the Jinchuan intrusion, adding complexity to the genesis of sulfide mineralization. We measured chalcophile elements and collected exploration data on Cu and Ni content to investigate their genesis and prospecting potential. The lower fine-grained subunits, comprising a layer of disseminated sulfides (DS-1) in the bottom and overlain by massive sulfide (MS) and olivine-sulfide cumulates (NTS-1), were overlain by the upper fine-grained subunits disseminated sulfides (DS-2) with a core of net-textured sulfides (NTS-2). The coarse-grained lherzolite with disseminated sulfides (DS-3) sharply cut through the upper subunit. All disseminated samples have Cu/Pd ratios (11,414–128,626) that exceed the mantle range. Additionally, the Pd/Ru ratios of DS-1 (1.26–13.2) are the lowest, followed by DS-2 (3.69–13.5), and highest in DS-3 (2.80–33.6). The exploration data indicate that the Ni and Cu contents and Cu/(Cu + Ni) ratios are significantly higher in DS-3 than in DS-2, while DS-1 shows more dispersed. The NTS-1 exhibits significant Ir and Ru depletion, with Cu/(Cu + Ni) decreasing with depth, eventually encountering Ir and Ru-riched MS. In addition, NTS-2 shows partially Ir depletion but Ru non-depletion, and Cu/(Cu + Ni) increases with depth. These signatures suggest that the lower and upper subunits, along with the coarse-grained unit, formed from platinum-group elements (PGE) in increasingly depleted magma, likely resulting from prior sulfide segregation before emplacement. Furthermore, the differences in Ir and Ru depletion and the variation in the Cu/(Cu + Ni) trend with depth between NTS-1 and NTS-2 were explained by the migration direction of fractional sulfide liquid during sulfide fractionation. We propose that the western end of the Jinchuan intrusion was formed by multi-stage magma emplacement. The strong correlation between exploration data (Ni, Cu, and Cu/(Cu + Ni)) and various sulfide mineralization layers suggests that delineating the spatial range of each sulfide mineralization could provide valuable information for deep mineral prospectivity mapping.

MGCoT: Multi-Grained Contextual Transformer for table-based text generation

Recent advances in Transformer have led to the revolution of table-based text generation. However, most existing Transformer-based architectures ignore the rich contexts among input tokens distributed in multi-level units (e.g., cell, row, or column), leading to sometimes unfaithful text generation that fails to establish accurate association relationships and misses vital information. In this paper, we propose Multi-Grained Contextual Transformer (MGCoT), a novel architecture that fully capitalizes on the multi-grained contexts among input tokens and thus strengthens the capacity of table-based text generation. The key primitive, Multi-Grained Contexts (MGCo) module, involves two components: a local context sub-module that adaptively gathers neighboring tokens to form the token-wise local context features, and a global context sub-module that consistently aggregates tokens from a broader range to form the shared global context feature. The former aims at modeling the short-range dependencies that reflect the salience of tokens within similar fine-grained units (e.g., cell and row) attending to the query token, while the latter aims at capturing the long-range dependencies that reflect the significance of each token within similar coarse-grained units (e.g., multiple rows or columns). Based on the fused multi-grained contexts, MGCoT can flexibly and holistically model the content of a table across multi-level structures. On three benchmark datasets, ToTTo, FeTaQA, and Tablesum, MGCoT outperforms strong baselines by a large margin on the quality of the generated texts, demonstrating the effectiveness of multi-grained context modeling. Our source codes are available at https://github.com/Cedric-Mo/MGCoT.

Clause-aware extractive summarization with topical decoupled contrastive learning

The sentence-level extracted summary is inevitably mixed with redundant information due to the uninformative phrases or detailed expressions mixed in it. The extraction of fine-grained units is dedicated to retain the semantical integrity. To keep the balance between text redundancy and semantical integrity, we propose a novel clause-aware summarization model (TDCL-ClauseSum). We separate complex sentences into grammatically independent but semantically dependent clauses. The clause is regarded as the extraction unit and leverage graph neural network and topical information to capture clause-level relationship. Then a decoupled contrastive loss is stacked over the neural model to fill the gap between topic prediction and clause classification. The experiments of TDCL-ClauseSum are evaluated on two public benchmark datasets CNN/daily mail and New York Times, which contain 310574 and 150536 samples, respectively. Various experiments show that our method achieves remarkable performance on the two datasets (CNN/daily mail:43.94/20.65/40.75, New York Times:49.69/29.84/43.01, in ROUGE-1/ROUGE-2/ROUGE-L). Its promising performance demonstrates that the superiority of clause extraction.

Basin-wide unconventional potential of the Doig Formation, Triassic of the Western Canada Sedimentary Basin

The Doig Formation is a Lower to Middle Triassic fine-grained unit of the Western Canada Sedimentary Basin, which includes a phosphate-bearing basal unit informally defined as the Doig Phosphate Zone (DPZ). The DPZ is recognized as a significant source of hydrocarbons hosted in conventional reservoirs in the basin. Phosphatic mudstones are uncommon and studies of their source-rock potential are scarce. The regional distribution of the Doig Formation source-rock properties across its entire stratigraphic and spatial range is poorly understood. In this study, we characterize the source-rock potential of the Doig Formation through an integrated assessment of thickness, organic content, kerogen type and thermal maturity. A new method for estimating the relative contribution of kerogen type and its original hydrogen index pseudo-van Krevelen, is introduced. The method assumes a binary mixture of kerogen types and requires an assumption for the initial value of immature kerogen from each end-member. The Doig Formation thickens southwestwards from the eastern erosional edge towards the deformation front to a maximum thickness of 229 m. The upper Doig and DPZ have distinct thickness distribution patterns, which suggests a transition from a tectonically-influenced sedimentation marked by uplifting, faulting and erosion, to the start of physiographic healing and a return to large-scale subsidence. The source-rock generative potential of the Doig Formation is fair to good, containing kerogen Types II and III. The DPZ has higher total organic carbon (TOC) content, of up to 12.2% and a higher proportion of Type II kerogen than that the upper Doig, where the maximum TOC content is 8.4 wt%. Original TOC was higher in the northern part of the lateral extent, which is likely related to a connection with open ocean; and in a linear trend through the center of the basin, which represents the trough of a marine fore-arc basin. Thermal maturity increases to the southwest, from early oil to dry gas windows, in a trend that approximately correlates to gas wetness and liquid yield. A zone of dry gas occurs near the eastern edge, however, which may be explained by preferential updip migration of thermogenic methane generated at greater depths, or mixing of thermogenic and biogenic gas.

Hydrovolcanic eruptions of the Paraná Igneous Province: Insights from mafic volcaniclastic deposits in Sertanópolis, Paraná, Brazil

In many continental flood basalt provinces (CFBPs), primary volcaniclastic deposits formed by explosive hydrovolcanism have been shown to be significant counterparts of better-known effusive facies, a notion that challenged traditional views of CFB evolution. For a long time, the only exception to this was the Paraná Igneous Province in South America. In order to better characterize and understand this setting, detailed geological mapping and petrographic studies were carried out on interflow beds at Sertanópolis, a 250 km2 region in northwest Paraná. Four interflow units were identified at the transition between the high and low TiO2 volcanic sequences. Fine-grained interflow facies include vesiculated clastic rocks, with structures ranging from diffuse to well defined stratification, decimeter-scale dune-bedding with preserved backset and foreset laminations. They are formed mainly by fine accessory (non-juvenile) siliciclastic ash (20–70%) and extremely fine ash or mud (10–50%), followed by fine hydrothermal mica (<5–40%), commonly hydrated (clay-mineral-rich) ash aggregates (<5–20%) and lesser amounts of juvenile and accessory glass and basalt fragments (ash to block-sized, <5–30%). In the eastern portion of the research area, a 4 × 2.5 km region with outcrops of scoriaceous basaltic breccias was mapped. These are mainly matrix to clast-supported breccias with finely laminated matrix similar to the fine-grained units. These rocks are interpreted as accessory-rich mafic pyroclastic beds, and their features suggest that multiple pyroclastic density currents (PDCs) formed each interflow unit, with minor discrete fallout contribution in medial successions. Pyroclastic breccias represent proximal PDC and ballistic deposits that help outline a possible conduit in which magmatic-like fragments are dominant in the framework. Stratified tuff-breccia strata in medial deposits represent progressive dilution of PDCs and continuous reworking of scoriaceous fragments by multiple currents. The local stratigraphy is consistent with repeated explosive hydrovolcanic eruptions in a maar-diatreme-like system similar to hydromagmatic volcanoes of other CFBPs.

Formation Processes of Gully-side Debris-Cones Determined from Ground-Penetrating Radar (Mt. Unzen, Japan)

Volcanic gullies develop U-shape cross-profiles due to the lahars that shape them after eruptive events. Over time, the processes of sediment-transfer change and the erosion of sub-vertical walls become a leading process. It results in debris-cones at their foot. However, the processes that generate these landforms is still unclear. Indeed, the surface is the one of a cone or a “steep debris-fan”, but are these diffusive landforms starting through progressive regressive-erosion of the wall, or do they follow an original collapse? This we do not know, and as the material is poorly-sorted with a mixture of unconsolidated fine sands to large clasts, digging a trench is not possible as it destroys the material structure. Therefore, the present contribution aims to define the development mechanisms of the side-wall debris-cones using Ground Penetrating Radar imaging, working in the Tansandani Gully at Unzen Volcano (Japan). The GPR used in the study is a Mala Ramac Pro-Ex mounted with a 500 MHz shielded antenna. The GPR signal penetrated up to 4–5 m depth and the velocity was homogeneous on average through the different material layers and at the different location, although at depth variations occurred. The structure of the units in the radargrams present units' parallel to the surface for the first 40 to 60 cm, with underneath the presence of more irregular prograding units, with eventually underneath a set of units that are structureless and often rich in blocks. The two lower structureless-units are not always present and one of the debris-cone only showed a set of rough subparallel units. This differences are the result of the debris-cones being generated by (1) at first a wall collapse or by rapid water-borne processes creating (2) bulk prograding units, which have to be put in place during a heavy-rainfall event. Finally, near the surface, the (3) thin subparallel units are most probably post-rainfall deposition, resulting from gravity process in dry-material. This can be inferred because there is no evidence of erosion or remobilization in layers and the process has been observed in the field. For the observed debris-cones, their formation is dominated by a wall collapse, which is then recovered by finer-grained units deposited sub-horizontally to the surface.

VPGB: A granular-ball based model for attribute reduction and classification with label noise

Neighborhood rough set (NRS) is an important tool for granular computing. It can handle discrete and continuous data without a prior discretization. However, the definitions of NRS are so strict that most algorithms based on NRS perform poorly on datasets with label noise. Existing classifiers mainly learn using the most fine-grained unit, that is, the sample or pixel point; thus, they have limited robustness to label noise. Moreover, few models can deal with both attribute reduction and classification with label noise. Granular-ball computing (GBC) is a novel granular computing model with a significant advantage in terms of robustness. Inspired by GBC, a robust variable parameter granular-ball model (VPGB) is proposed in this study to achieve both attribute reduction and classification in a label noise environment from a coarse granularity perspective. In VPGB attribute reduction, first, the related notions of NRS are defined for the label noise environment. Second, the universe is granulated into granular-balls of the same size (called FGBs), with the centers and radii being fixed. Third, attribute reduction is performed based on the FGBs. In VPGB classification, the universe is granulated into adaptive granular-balls of different sizes (called AGBs), with adaptive centers and radii. Then, the test samples are predicted based on the distance between the test samples and AGBs. The experimental results on public datasets show that the VPGB model performs better in accuracy and robustness than the current state-of-the-art algorithms in terms of both attribute reduction and classification.

Detecting fake news by enhanced text representation with multi-EDU-structure awareness

Since fake news poses a serious threat to society and individuals, numerous studies have been brought by considering text, propagation and user profiles. Due to the data collection problem, these methods based on propagation and user profiles are less applicable in the early stages. A good alternative method is to detect news based on text as soon as they are released, and a lot of text-based methods were proposed, which usually utilized words, sentences or paragraphs as basic units. But, word is a too fine-grained unit to express coherent information well, sentence or paragraph is too coarse to show specific information. Which granularity is better and how to utilize it to enhance text representation for fake news detection are two key problems. In this paper, we introduce Elementary Discourse Unit (EDU) whose granularity is between word and sentence, and propose a multi-EDU-structure awareness model to improve text representation for fake news detection, namely EDU4FD. For the multi-EDU-structure awareness, we build the sequence-based EDU representations and the graph-based EDU representations. The former is gotten by modeling the coherence between consecutive EDUs with TextCNN that reflect the semantic coherence. For the latter, we first extract rhetorical relations to build the EDU dependency graph, which can show the global narrative logic and help deliver the main idea truthfully. Then a Relation Graph Attention Network (RGAT) is set to get the graph-based EDU representation. Finally, the two EDU representations are incorporated as the enhanced text representation for fake news detection, using a gated recursive unit combined with a global attention mechanism. Experiments on four cross-source fake news datasets show that our model outperforms the state-of-the-art text-based methods. Our results suggest that considering EDU and its structural features could enhance the text representation for fake news detection.

Lithologic and Geochemical Constraints on the Genesis of a Newly Discovered Orebody in the Jinchuan Intrusion, NW China

AbstractThe Jinchuan Ni-Cu-platinum group element (PGE) sulfide deposit is one of the world’s major magmatic sulfide deposits. The Jinchuan intrusion originally consists of two independent bodies, named the Western and Eastern intrusions. Recently, an underground exploration program discovered a large economic orebody in the Fine-Grained unit of the Western intrusion, independent of the orebody No. 24 that occurs at the base of the Coarse-Grained unit. The newly discovered orebody dips to the southwest with thicknesses of 20 to 150 m and is composed of olivine-sulfide cumulates with net-textured sulfide overlain by lherzolite with disseminated sulfide. The net-textured olivine-sulfide cumulates have average grades of 2.1 wt % Ni and 1.2 wt % Cu, whereas the disseminated ores in the lherzolite have average grades of 0.6 wt % Ni and 0.4 wt % Cu. This makes this orebody, named orebody No. 3 and hosted in segment III, the fourth largest at Jinchuan. The sharp lithologic and chemical variation between the net-textured olivine-sulfide cumulates and the disseminated sulfide-bearing lherzolite indicates that they were formed by different processes. Evident differences in sulfidefree whole-rock compositions reflect the dunite forming as olivine adcumulate and the lherzolite as olivine orthocumulate with variable trapped liquid abundances. The olivine-sulfide cumulates were formed by percolation of sulfide liquid into the olivine adcumulates, expelling the interstitial silicate melts upward, whereas the disseminated ores in the lherzolite were formed by settling of finely dispersed sulfide droplets in the olivine crystal mush, within which pyroxene crystallized from the interstitial melts. Variations of PGE tenors of the net-textured sulfides demonstrate fractionation of the sulfide liquids. The relatively high PGE tenors of the disseminated sulfides compared with the net-textured sulfides are associated with upgrading of the sulfides due to reaction with fresh magma.

Stratigraphic and structural controls on groundwater salinity variations in the Poso Creek Oil Field, Kern County, California, USA

Groundwater total dissolved solids (TDS) distribution was mapped with a three-dimensional (3D) model, and it was found that TDS variability is largely controlled by stratigraphy and geologic structure. General TDS patterns in the San Joaquin Valley of California (USA) are attributed to predominantly connate water composition and large-scale recharge from the adjacent Sierra Nevada. However, in smaller areas, stratigraphy and faulting play an important role in controlling TDS. Here, the relationship of stratigraphy and structure to TDS concentration was examined at Poso Creek Oil Field, Kern County, California. The TDS model was constructed using produced water TDS samples and borehole geophysics. The model was used to predict TDS concentration at discrete locations in 3D space and used a Gaussian process to interpolate TDS over a volume. In the overlying aquifer, TDS is typically <1,000 mg/L and increases with depth to ~1,200–3,500 mg/L in the hydrocarbon zone below the Macoma claystone—a regionally extensive, fine-grained unit—and reaches ~7,000 mg/L in isolated places. The Macoma claystone creates a vertical TDS gradient in the west where it is thickest, but control decreases to the east where it pinches out and allows freshwater recharge. Previously mapped normal faults were found to exhibit inconsistent control on TDS. In one case, high-density faulting appears to prevent recharge from flushing higher-TDS connate water. Elsewhere, the high-throw segments of a normal fault exhibit variable behavior, in places blocking lower-TDS recharge and in other cases allowing flushing. Importantly, faults apparently have differential control on oil and groundwater.

Long-term environmental changes in the Geum Estuary (South Korea): Implications of river impoundments

We investigated a sediment core collected from the Geum Estuary through sedimentological and geochemical analyses. Three lithological units were classified based on sedimentological characteristics. Unit 1 and Unit 3 were geochemically distinct, while Unit 2 was the transitional phase between them. The geochemical results suggest that the contribution of terrestrial organic carbon (OC) to the sedimentary OC pool in the coarse-grained Unit 1 was lesser than that of fine-grained Unit 3. The excess activity (210Pbex) and the sedimentation rate indicate that Unit 1 corresponded to 1977 Common Era (CE). Since the first dam construction on the Geum River began in 1975 CE, the deposition of Unit 1 in the Geum Estuary is likely associated with river impoundments, which reduce the delivery of fine-grained sediment and terrestrial OC to the estuary. This study highlights the role of river impoundments in altering the sedimentary OC and thus the sedimentary environment in the estuary.

Evidence for geologically recent explosive volcanism in Elysium Planitia, Mars

Volcanic activity on Mars peaked during the Noachian and Hesperian periods but has continued since then in isolated locales. Elysium Planitia hosts numerous young, fissure-fed flood lavas with ages ranging from approximately 500 to 2.5 million years (Ma). We present evidence for a fine-grained unit that is atypical of aeolian deposits in the region and may be the youngest volcanic deposit yet documented on Mars. The unit has a low albedo, high thermal inertia, includes high‑calcium pyroxene-rich material, and is distributed symmetrically around a segment of the Cerberus Fossae fissure system in Elysium Planitia. This deposit is superficially similar to features interpreted as pyroclastic deposits on the Moon and Mercury. Unlike previously documented lava flows in Elysium Planitia, this feature is morphologically consistent with a fissure-fed pyroclastic deposit, mantling the surrounding lava flows with a thickness on the order of tens of cm over most of the deposit and a volume of 1.1–2.8 × 107 m3. Thickness and volume estimates are consistent with tephra deposits on Earth. Stratigraphic relationships indicate a relative age younger than the surrounding volcanic plains and the Zunil impact crater (~0.1–1 Ma), with crater counting suggesting that the deposit has an absolute model age of 53 ± 7 to 210 ± 12 ka. This young age implies that if this deposit is volcanic then the Cerberus Fossae region may not be extinct and that Mars may still be volcanically active. This interpretation is consistent with the identification of seismicity in this region by the Interior Exploration using Seismic Investigations, Geodesy, and Heat Transport (InSight) lander, and has additional implications for astrobiology.

Characterizing Fine-Grained Resource Utilization for Multitasking GPGPU in Cloud Systems

Managing GPGPU resources in cloud systems is challenging as workloads with various resource usage patterns coexist. To determine the co-location of workloads, previous studies have shown that run-time performance profiling and dynamic relocation of workloads is necessary due to interference between workloads. However, this makes instant scheduling difficult and also affects the performance of workload executions. In this article, we show that efficient resource sharing in GPGPU is possible without run-time profiling if resource usage characteristics of workloads are analyzed down to a fine-grained unit level. To extract workload characteristics, we do not perform profiling at scheduling time, but separates profiling from scheduling, thereby reducing the run-time complexity of previous approaches. Specifically, we anatomize the characteristics of various GPGPU workloads and present a new scheduling policy that aims at balancing resource utilization by co-locating workloads with complementary resource demands. Simulation experiments under various virtual machine scenarios show that the proposed policy improves the GPGPU throughput by 119.5% on average and up to 191.7%.

Evaluation of Land Consolidation and Renovation Processin the Decayed Textures: A Case of Study the Nezamabad Neighborhood in Tehran

Decayed textures have been one of the most important issues and challenges of Tehran's urban management in the last two decades. Therefore, numerous and varied policies, programs, and actions have been adopted and implemented in these decayed areas. The land consolidation is one of the approaches of the renovation of the old and decayed urban texture of Tehran based on resolution 300/310/15301 dated 24/07/2006 of the Supreme Council of Urban Planning and Architecture of Iran and clauses 9-5 of the detailed design rules and regulations finely executed. The purpose of this study was to evaluate the process of land consolidation in the renovation of the urban decayed texture of the Nizamabad neighborhood in zone 1, district 7 of Tehran Municipality. Most of the information required for this research was collected by the field study method. The respondent sampling of this study consisted of the developers, experts, fine-grained unit owners of the combined fine-grained parts of this neighborhood. SPSS software was used to analyze the descriptive data. Chi-Square and Friedman nonparametric statistical tests were used depending on the type of data, the scale of variables and goals. According to the results of this study, the process of implementation of consolidation renovation in the Nizamabad neighborhood has been faced with numerous and various damages.

The origin of carbonate cements in the Hildasay reservoir, Cambo Field, Faroe-Shetland Basin; clumped isotopic analysis and implications for reservoir performance

The early Eocene paralic sandstones of the Hildasay Member of the Flett Formation form the major oil-bearing reservoir in the Cambo Field, located in the Faroe-Shetland Basin. The sandstones locally contain calcite-cemented intervals that vary in thickness from decimetre to over 1 m. A calcite-cemented interval from well 204/10a-5 has been analysed petrographically and using clumped isotopes to determine its mode of formation and potential lateral extent. Petrographic analysis shows the cemented interval to consist of ferroan calcite, with a consistent dull red cathodoluminescence, suggesting a single phase of precipitation. The centre of the cemented interval comprises a finer grained unit with detrital kerogen and early sphaerosiderite, while the rest comprises homogeneous porosity-occluding ferroan calcite. The early sphaerosiderite in the centre is replaced by ferroan calcite with a high δ13 carbon isotopic signature (δ13 CVPDB = 7.53‰) suggesting that it formed during anaerobic methanogenesis or fermentation. Samples from the intermediate zone have a lower δ13 carbon isotopic composition (δ13 CVPDB = 0.72 to −3.68‰). In comparison the outer margin of the cemented unit has an even lower δ13 carbon isotopic signature (δ13 CVPDB = −15.5 to -15.9‰) more typical of a strong aerobic oxidation source. The oxygen isotopic signature of the cements is similar (δ18 OVPDB = −10.9 to −12.2‰). Analysis of the clumped isotopes suggest that the ferroan calcite formed at ∼40–50 °C, from pore waters that were predominantly meteoric in origin (δ18 OVSMOW = −4.8 to −5.7‰). Burial history modelling would suggest that the cements formed at a depth of ∼500–1000 m. Meteoric water was likely to have been introduced during the formation of the mid-Lutetian unconformity ∼45Mya, approximately 10 Mya after the sandstones were deposited. The model proposed is that ferroan calcite precipitation was initiated in the fine lag deposits that contained the kerogen and sphaerosiderite, and then grew outwards. If the model is correct, the cemented units should be restricted to this facies and consequently of limited lateral extent. Consequently, it is likely that the cemented intervals will have a limited impact upon the reservoir performance and are unlikely to act as major barriers to fluid flow.

Can Fine-Grained Functional Split Benefit to the Converged Optical-Wireless Access Networks in 5G and Beyond?

The centralized radio access network (C-RAN) is an effective architecture to promote CAPEX/OPEX reduction and cell cooperation derived from its centralized baseband processing. However, there is a contradiction between centralization gain and transport resource saving, which hinders the vision of a resource-efficient and cost-effective RAN deployment. Advanced RAN architectures with functional splits are then introduced to cope with this challenge. Distinguished with other studies, we are intended to investigate whether a fine-grained functional split architecture could benefit to the RAN evolution, and how it impacts on the converged optical-wireless access networks. To this end, we establish a quantitative model to analyze the performance of this architecture. With the fine-grained split, baseband unit (BBU) is divided into a set of fine-grained units (FU) to be placed in desired processing pools (PP) as a service chain. To analyze the placement performance, we propose a mixed-integer linear programming model (MILP) considering the PP selection, routing, wavelength and bandwidth assignment, as well as latency control to minimize the number of PPs, bandwidth, latency, and functions deployment cost. We compare its performance with other two coarse-grained split architectures, i.e., SBBU (adopt low-PHY split like BBU in 4G) and recently emerged DU-CU in both small-scale and large-scale network scenarios. Our analyses provide insights into the modeling and design of efficient converged optical-wireless access networks in 5G and beyond.

Integrating lexical and prosodic features for automatic paragraph segmentation

Spoken documents, such as podcasts or lectures, are a growing presence in everyday life. Being able to automatically identify their discourse structure is an important step to understanding what a spoken document is about. Moreover, finer-grained units, such as paragraphs, are highly desirable for presenting and analyzing spoken content. However, little work has been done on discourse based speech segmentation below the level of broad topics. In order to examine how discourse transitions are cued in speech, we investigate automatic paragraph segmentation of TED talks using lexical and prosodic features. Experiments using Support Vector Machines, AdaBoost, and Neural Networks show that models using supra-sentential prosodic features and induced cue words perform better than those based on the type of lexical cohesion measures often used in broad topic segmentation. Moreover, combining a wide range of individually weak lexical and prosodic predictors improves performance, and modelling contextual information using recurrent neural networks outperforms other approaches by a large margin. Our best results come from using late fusion methods that integrate representations generated by separate lexical and prosodic models while allowing interactions between these features streams rather than treating them as independent information sources. Application to ASR outputs shows that adding prosodic features, particularly using late fusion, can significantly ameliorate decreases in performance due to transcription errors.

Gas hydrate quantification in Walker Ridge block 313, Gulf of Mexico, from full-waveform inversion of ocean-bottom seismic data

The Gulf of Mexico (GOM) Joint Industry Project Leg 2 logging-while-drilling data in Walker Ridge lease block 313 (WR313) in the GOM detected gas hydrate in coarse- and fine-grained sediments at sites WR313-G and WR313-H. The coarse-grained units are thin ([Formula: see text]) and highly saturated, whereas the fine-grained unit is thick (approximately 200 m) with low saturation and fracture-filling gas hydrate. Unlike its coarse-grained counterpart, the seismic character of the fine-grained unit does not clearly indicate the presence of gas hydrate, which would likely have remained undiscovered in the absence of drilling. In this paper, through frequency-domain acoustic full-waveform inversion (FWI) of ocean-bottom seismometer data along a 2D multichannel seismic transect near sites WR313-G and WR313-H, we detect and quantify gas hydrate in the fine-grained unit. Key results are as follows: First, the base of the gas hydrate stability zone, which is not obvious in the reflection profile, can be discerned in the FWI results. Second, the gas hydrate in the fine-grained unit is mainly confined to the area between two sets of opposite-dipping normal faults implying that the fault architecture may be partially responsible for this gas hydrate accumulation and distribution.

Hierarchical parameterization and compression-based object modelling of high net:gross but poorly amalgamated deep-water lobe deposits

Deep-water lobe deposits are arranged hierarchically and can be characterized by high net:gross ratios but poor sand connectivity due to thin, but laterally extensive, shale layers. This heterogeneity makes them difficult to represent in standard full-field object-based models, since the sands in an object-based model are not stacked compensationally and become connected at a low net:gross ratio. The compression algorithm allows the generation of low-connectivity object-based models at high net:gross ratios, by including the net:gross and amalgamation ratios as independent input parameters. Object-based modelling constrained by the compression algorithm has been included in a recursive workflow, permitting the generation of realistic models of hierarchical lobe deposits. Representative dimensional and stacking parameters collected at four different hierarchical levels have been used to constrain a 250 m-thick, 14 km 2 model that includes hierarchical elements ranging from 20 cm-thick sand beds to more than 30 m-thick lobe complexes. Sand beds and the fine-grained units are represented explicitly in the model, and the characteristic facies associations often used to parameterize lobe deposits are emergent from the modelling process. The model is subsequently resampled without loss of accuracy for flow simulation, and results show clearly the influence of the hierarchical heterogeneity on drainage and sweep efficiency during a water-flood simulation.

The Bouse Formation, a controversial Neogene archive of the evolving Colorado River: a scientific drilling workshop report (28 February–3 March 2019 – BlueWater Resort &amp; Casino, Parker, AZ, USA)

Abstract. Neogene deposits of the lower Colorado River valley, especially the Miocene(?) and early Pliocene Bouse Formation, have been the focus of intense debate regarding the early paleoenvironmental history of this important continental-scale river system in southwestern North America and its integration with the proto-Gulf of California. Fine-grained units within these Neogene deposits also hold a promising archive of Pliocene paleoclimate history for this part of the world. Because the depocenter deposits of the Bouse Formation and the deposits that overlie and underlie it are poorly exposed and highly weathered, the formation is ripe for study through collection of drill cores. A workshop was held 28 February–3 March 2019 in Parker, AZ, USA, to discuss how scientific drilling might be employed to help resolve the Bouse controversies and improve our understanding of paleoclimate history in the region.