Exploiting Positional Information for Session-Based Recommendation

For present e-commerce platforms, it is important to accurately predict users’ preference for a timely next-item recommendation. To achieve this goal, session-based recommender systems are developed, which are based on a sequence of the most recent user-item interactions to avoid the influence raised from outdated historical records. Although a session can usually reflect a user’s current preference, a local shift of the user’s intention within the session may still exist. Specifically, the interactions that take place in the early positions within a session generally indicate the user’s initial intention, while later interactions are more likely to represent the latest intention. Such positional information has been rarely considered in existing methods, which restricts their ability to capture the significance of interactions at different positions. To thoroughly exploit the positional information within a session, a theoretical framework is developed in this paper to provide an in-depth analysis of the positional information. We formally define the properties of forward-awareness and backward-awareness to evaluate the ability of positional encoding schemes in capturing the initial and the latest intention. According to our analysis, existing positional encoding schemes are generally forward-aware only, which can hardly represent the dynamics of the intention in a session. To enhance the positional encoding scheme for the session-based recommendation, a dual positional encoding (DPE) is proposed to account for both forward-awareness and backward-awareness . Based on DPE, we propose a novel Positional Recommender (PosRec) model with a well-designed Position-aware Gated Graph Neural Network module to fully exploit the positional information for session-based recommendation tasks. Extensive experiments are conducted on two e-commerce benchmark datasets, Yoochoose and Diginetica and the experimental results show the superiority of the PosRec by comparing it with the state-of-the-art session-based recommender models.

Analytic Functions in Local Shift-Invariant Spaces and Analytic Limits of Level Dependent Subdivision

In this paper we characterize all subspaces of analytic functions in finitely generated shift-invariant spaces with compactly supported generators and provide explicit descriptions of their elements. We illustrate the differences between our characterizations and Strang-Fix or zero conditions on several examples. Consequently, we depict the analytic functions generated by scalar or vector subdivision with masks of bounded and unbounded support. In particular, we prove that exponential polynomials are indeed the only analytic limits of level dependent scalar subdivision schemes with finitely supported masks.

Shear-wave seismic reflection processing - the importance of velocity analysis

<p>Based on our experience, one of the most important steps in processing shear-wave seismic reflection data is the velocity analysis. In unconsolidated materials a very fine velocity analysis is more essential for S-waves than for P-waves because shear-wave velocities vary over several orders of magnitude and can change very quickly laterally and with depth. Velocities between 100m/s in glaciolacustine/marine deposits (clay-sized silts) and 1200m/s in stiff diamicton (till) were encountered in recent surveys. Shear-wave velocities have the large advantage of not being changed by the phase of the pore content such as the groundwater table.</p><p>We present two fundamentally different methods for velocity determination: 1) velocity semblance analysis based on hyperbolic reflection move-out on common midpoint (cmp) gathers and 2) Local Phase – Local Shift (LPLS) method which automatically estimates the reflection slope (local static shift) in the time-frequency domain of cmp gathers. Published in 2020, the latter method can be used for automated processing and substantially saves processing time.</p><p>Processing steps in preparation for velocity analysis (independent of the chosen method) include frequency filtering, trace equalizing and muting. We show velocity semblance images from different geological settings (glacial, postglacial) and from different shear components and discuss differences. Information gained besides shear velocities include mapped reflectors and located diffractions. Using those examples, we demonstrate how combining all information using visualisation techniques enhances interpretation of such data sets.</p>

Snail assemblages in Holocene floodplain research – an example from the southern Caucasus

During the last decades, rivers and their deposits in different regions were intensively studied to better understand the late-Quaternary landscape evolution and former human activities. One proxy for paleoecological and paleoclimatic reconstructions is the analysis of gastropods (snails) from carbonatic river sediments. In the scope of this study, we investigated gastropod assemblages from a Holocene fluvial sediment–paleosol sequence at the upper Alazani River in the southeastern Caucasus. On the one hand, we aimed to derive reliable independent information about possible long-lasting human activity since the late Neolithic–Chalcolithic in the upper Alazani floodplain. This was formerly suggested by n-alkane biomarker vegetation reconstructions from the fluvial sediments. However, the reliability of that method is still debated. On the other hand, we aimed to obtain reliable information about a formerly suggested, possibly tectonic-driven, large-scale shift of the river course during the Late Holocene. In agreement with the n-alkane biomarkers, our results demonstrate that the studied site was free of the natural forests during the Early and Middle Holocene until ca. 4.5 cal kyr BP. Since this contrasts with a pollen-based vegetation reconstruction from a neighboring floodplain that was covered with forests during that time, the open vegetation in the upper Alazani valley was probably caused by continuous settlement activity as is also indicated by archeological finds in the sequence. Therefore, using our paleoecological proxies it is possible that we identified a settlement center in the upper Alazani floodplain that was populated from the late Neolithic–Chalcolithic. This center was not known thus far, since the settlement remains are covered by thick floodloam today that hindered their detection during archeological surface surveys. Therefore, our findings suggest that the area in the southern Caucasus region that was settled during the late-Neolithic–Chalcolithic period should have been larger than was known thus far. Furthermore, increasing contributions of wetland gastropods since ca. 4 cal kyr BP confirm a local shift of the river course towards the investigated site during the Late Holocene, possibly linked with ongoing tectonic activity. However, in contrast to former suggestions our gastropod assemblages indicate a slow rather than abrupt process. Our study demonstrates the high value of gastropod assemblages for geoarcheological and geomorphological research in floodplains with carbonatic river sediments at both a regional and local spatial scale.

Localisation of gold ores in faults of different types

The results of the author’s and general works of domestic and foreign geologists, who studied the location of the largest gold deposits in fault zones, characterised by the structure of ore-bearing zones and the concentration of reserves of ores and gold of different scale in them, are presented. The main reasons for such differences are considered. The longest faults on our planet are regional shifts. They are continuously traced for hundreds (up to 1,400) of kilometres along the boundaries of gold-bearing belts and provinces. However, gold ore deposits are located in their zones at extremely limited (point) intervals not exceeding 3—5 km. They are always enclosed between ancient transverse or oblique-oriented fractures of deep, most likely mantle, formation and penetration. In all mineralised faults, gold ore bodies are localised in various geological and structural traps, which are considered in the article and are reflected in the plans and sections. The crossing nodes of regional shifts, as well as overfaults and faults of transverse faults (and dislocations), act as the main promising objects in the deposits search and exploration. Obviously, such nodes should be considered as direct signs of the possible evidence of gold ores. The internal structure of the world leader, gold-bearing Muruntaussky (North-East) local shift (Uzbekistan), studied in detail by the author and other geologists, is given as an example.

Effect of Severe Plastic Deformation on the Structure and Properties of the Aluminum Alloy System Al-Cu-Mg

The structure and properties of discs from cast coarse-grained D16 alloy have been investigated with the help of optical metallography and transmission electron microscopy, after the traditional heat treatment (HT) modes or subjected to high pressure torsion (HPT) under pressure in 6 GPa at room temperature. Standard HT modes included: Т4 (hardening 495°С + natural aging at room temperature for 5 days) and Т6 (hardening 495°С + artificial aging at 185°С for 10 hours). It is shown that after HT of alloy D16 according to modes T6 and Т4, the sizes of the structural elements, compared with the original, decreased by about 4,8 times and the microhardness increased by 1,6 times. It is shown that after one or ten rotates of HPT in D16 alloy the sizes of structural elements, in comparison with initial, decreased by 393 and 899 times, and microhardness increased by 2,4 and 2,9 times, respectively. Detection of hardening hard, brittle and high-alloy copper γ2- phases (Cu9Al4) indicates the passage in the HPT process of the decomposition of a supersaturated solid solution (dynamic aging) in the planes of a local shift.

Edge state modulation by mean viscosity gradients

Motivated by the relevance of edge state solutions as mediators of transition, we use direct numerical simulations to study the effect of spatially non-uniform viscosity on their energy and stability in minimal channel flows. What we seek is a theoretical support rooted in a fully nonlinear framework that explains the modified threshold for transition to turbulence in flows with temperature-dependent viscosity. Consistently over a range of subcritical Reynolds numbers, we find that decreasing viscosity away from the walls weakens the streamwise streaks and the vortical structures responsible for their regeneration. The entire self-sustained cycle of the edge state is maintained on a lower kinetic energy level with a smaller driving force, compared to a flow with constant viscosity. Increasing viscosity away from the walls has the opposite effect. In both cases, the effect is proportional to the strength of the viscosity gradient. The results presented highlight a local shift in the state space of the position of the edge state relative to the laminar attractor with the consequent modulation of its basin of attraction in the proximity of the edge state and of the surrounding manifold. The implication is that the threshold for transition is reduced for perturbations evolving in the neighbourhood of the edge state in the case that viscosity decreases away from the walls, and vice versa.

Comparison of methods for extracting ADCIGs from RTM

Methods for extracting angle-domain common-image gathers (ADCIGs) during 2D reverse-time migration fall into three main categories; direction-vector-based methods, local-plane-wave decomposition methods, and local-shift imaging condition methods. The direction-vector-based methods, which use either amplitude gradients or phase gradients, cannot handle overlapping events because of an assumption of one propagation direction per imaging point per imaging time; however, the ADCIGs from the direction-vector-based methods have the highest angle resolution. A new direction-vector-based method using instantaneous phase gradients in space and time gives the same propagation directions and ADCIGs as those obtained by the Poynting vector or polarization vector based methods, where amplitudes are large. Angles calculated by the phase gradients have larger uncertainties at smaller amplitudes, but they do not significantly degrade the ADCIGs because they contribute only small amplitudes. The local-plane-wave decomposition and local-shift imaging condition methods, implemented either by a Fourier transform or by a slant stack transform, can handle overlapping events, and produce very similar angle gathers. ADCIGs from both methods depend on the local window size in which the transforms are done. In small local windows, both methods produce ADCIGs with low noise, but also with low angle resolution; in large windows, they have high angle resolution, but contain smeared artifacts.

Reconstruction of Atlantic water variability during the Holocene in the western Barents Sea

The gravity core JM09-KA11-GC from 345 m water depth on the western Barents Sea margin was investigated for down-core distribution patterns of benthic Foraminifera, stable isotopes, and sedimentological parameters in order to reconstruct the flow of Atlantic water during the Holocene. The core site is located below the Atlantic water masses flowing into the Arctic Ocean and close to the Arctic front. The results show continuous presence of Atlantic water at the margin throughout the Holocene. During the early Holocene, (11 500–9800 cal yr BP), bottom water temperatures as calculated by transfer functions rose by 1.5 °C, likely due to the increased inflow of Atlantic water, although sea ice was still present at this time. The transition to the mid-Holocene is characterized by a local shift in current regime, resulting in a ceased supply of fine-grained material to the core location. Throughout the mid-Holocene the δ18O values indicate a slight cooling, thereby following changes in insolation. In the last 1500 yr, inflow of Atlantic water increased but was interrupted by periods of increased influence of Arctic water causing periodically colder and more unstable conditions.