927 research outputs found
Recurrent Latent Variable Networks for Session-Based Recommendation
In this work, we attempt to ameliorate the impact of data sparsity in the
context of session-based recommendation. Specifically, we seek to devise a
machine learning mechanism capable of extracting subtle and complex underlying
temporal dynamics in the observed session data, so as to inform the
recommendation algorithm. To this end, we improve upon systems that utilize
deep learning techniques with recurrently connected units; we do so by adopting
concepts from the field of Bayesian statistics, namely variational inference.
Our proposed approach consists in treating the network recurrent units as
stochastic latent variables with a prior distribution imposed over them. On
this basis, we proceed to infer corresponding posteriors; these can be used for
prediction and recommendation generation, in a way that accounts for the
uncertainty in the available sparse training data. To allow for our approach to
easily scale to large real-world datasets, we perform inference under an
approximate amortized variational inference (AVI) setup, whereby the learned
posteriors are parameterized via (conventional) neural networks. We perform an
extensive experimental evaluation of our approach using challenging benchmark
datasets, and illustrate its superiority over existing state-of-the-art
techniques
Variational Deep Semantic Hashing for Text Documents
As the amount of textual data has been rapidly increasing over the past
decade, efficient similarity search methods have become a crucial component of
large-scale information retrieval systems. A popular strategy is to represent
original data samples by compact binary codes through hashing. A spectrum of
machine learning methods have been utilized, but they often lack expressiveness
and flexibility in modeling to learn effective representations. The recent
advances of deep learning in a wide range of applications has demonstrated its
capability to learn robust and powerful feature representations for complex
data. Especially, deep generative models naturally combine the expressiveness
of probabilistic generative models with the high capacity of deep neural
networks, which is very suitable for text modeling. However, little work has
leveraged the recent progress in deep learning for text hashing.
In this paper, we propose a series of novel deep document generative models
for text hashing. The first proposed model is unsupervised while the second one
is supervised by utilizing document labels/tags for hashing. The third model
further considers document-specific factors that affect the generation of
words. The probabilistic generative formulation of the proposed models provides
a principled framework for model extension, uncertainty estimation, simulation,
and interpretability. Based on variational inference and reparameterization,
the proposed models can be interpreted as encoder-decoder deep neural networks
and thus they are capable of learning complex nonlinear distributed
representations of the original documents. We conduct a comprehensive set of
experiments on four public testbeds. The experimental results have demonstrated
the effectiveness of the proposed supervised learning models for text hashing.Comment: 11 pages, 4 figure
The early Pliocene Titiokura Formation: stratigraphy of a thick, mixed carbonate-siliciclastic shelf succession in Hawke's Bay Basin, New Zealand
This paper presents a systematic stratigraphic description of the architecture of the early Pliocene Titiokura Formation (emended) in the Te Waka and Maungaharuru Ranges of western Hawke's Bay, and presents a facies, sequence stratigraphic, and paleoenvironmental analysis of the sedimentary succession. The Titiokura Formation is of early Pliocene (Opoitian-Waipipian) age, and unconformably overlies Mokonui Formation, which is a regressive late Miocene and early Pliocene (Kapitean to early Opoitian) succession. In the Te Waka Range and the southern parts of the Maungaharuru Range, the Titiokura Formation comprises a single limestone sheet 20-50 m thick, with calcareous sandstone parts. In the vicinity of Taraponui Trig, and to the northeast, the results of 1:50 000 mapping show that the limestone gradually partitions into five members, which thicken markedly to the northeast to total thicknesses of c. 730 m, and concomitantly become dominated by siliciclastic sandstone. The members (all new) from lower to upper are: Naumai Member, Te Rangi Member, Taraponui Member, Bellbird Bush Member, and Opouahi Member. The lower four members are inferred to each comprise an obliquity-controlled 41 000 yr 6th order sequence, and the Opouahi Member at least two such sequences. The sequences typically have the following architectural elements from bottom to top: disconformable sequence boundary that formed as a transgressive surface of erosion; thin transgressive systems tracts (TSTs) with onlap and backlap shellbeds, or alternatively, a single compound shellbed; downlap surface; and very thick (70-200 m) highstand (HST) and regressive systems tracts (RST) composed of fine sandstone. The sequences in the Opouahi Member have cryptic TSTs, sandy siltstone to silty sandstone HSTs, and cross-bedded, differentially cemented, fine sandstone RSTs; a separate variant is an 11 m thick bioclastic limestone (grainstone and packstone) at the top of the member that crops out in the vicinity of Lake Opouahi. Lithostratigraphic correlations along the crest of the ranges suggest that the Titiokura Formation, and its correlatives to the south around Puketitiri, represent a shoreline-to-shelf linked depositional system. Carbonate production was focused around a rocky seascape as the system onlapped basement in the south, with dispersal and deposition of the comminuted carbonate on an inner shelf to the north, which was devoid of siliciclastic sediment input. The rates of both subsidence and siliciclastic sediment flux increased rapidly to the northeast of the carbonate "platform", with active progradation and offlap of the depositional system into more axial parts of Hawke's Bay Basin
SchNet - a deep learning architecture for molecules and materials
Deep learning has led to a paradigm shift in artificial intelligence,
including web, text and image search, speech recognition, as well as
bioinformatics, with growing impact in chemical physics. Machine learning in
general and deep learning in particular is ideally suited for representing
quantum-mechanical interactions, enabling to model nonlinear potential-energy
surfaces or enhancing the exploration of chemical compound space. Here we
present the deep learning architecture SchNet that is specifically designed to
model atomistic systems by making use of continuous-filter convolutional
layers. We demonstrate the capabilities of SchNet by accurately predicting a
range of properties across chemical space for \emph{molecules and materials}
where our model learns chemically plausible embeddings of atom types across the
periodic table. Finally, we employ SchNet to predict potential-energy surfaces
and energy-conserving force fields for molecular dynamics simulations of small
molecules and perform an exemplary study of the quantum-mechanical properties
of C-fullerene that would have been infeasible with regular ab initio
molecular dynamics
Neural Attentive Session-based Recommendation
Given e-commerce scenarios that user profiles are invisible, session-based
recommendation is proposed to generate recommendation results from short
sessions. Previous work only considers the user's sequential behavior in the
current session, whereas the user's main purpose in the current session is not
emphasized. In this paper, we propose a novel neural networks framework, i.e.,
Neural Attentive Recommendation Machine (NARM), to tackle this problem.
Specifically, we explore a hybrid encoder with an attention mechanism to model
the user's sequential behavior and capture the user's main purpose in the
current session, which are combined as a unified session representation later.
We then compute the recommendation scores for each candidate item with a
bi-linear matching scheme based on this unified session representation. We
train NARM by jointly learning the item and session representations as well as
their matchings. We carried out extensive experiments on two benchmark
datasets. Our experimental results show that NARM outperforms state-of-the-art
baselines on both datasets. Furthermore, we also find that NARM achieves a
significant improvement on long sessions, which demonstrates its advantages in
modeling the user's sequential behavior and main purpose simultaneously.Comment: Proceedings of the 2017 ACM on Conference on Information and
Knowledge Management. arXiv admin note: text overlap with arXiv:1511.06939,
arXiv:1606.08117 by other author
Isolated Character Forms from Dated Syriac Manuscripts
This paper describes a set of hand-isolated character samples selected from securely dated manuscripts written in Syriac between 300 and 1300 C.E., which are being made available for research purposes. The collection can be used for a number of applications, including ground truth for character segmentation and form analysis for paleographical dating. Several applications based upon convolutional neural networks demonstrate the possibilities of the data set
Provenance and geochemistry of exotic clasts in conglomerates of the Oligocene Torehina Formation, Coromandel Peninsula, New Zealand
Non-marine pebble to cobble conglomerates of the lower Torehina Formation (Oligocene) crop out along western Coromandel Peninsula and overlie, with strong angular discordance, continental-margin metasedimentary rocks (Manaia Hill Group) of Mesozoic (Late Jurassic to ?Early Cretaceous) age. The conglomerates contain provenance information that identifies a pre-Oligocene depositional history obscured by the unconformable juxtaposition of these Tertiary and Mesozoic strata. Most clasts in the lower Torehina Formation are visually similar to local bedrock lithologies, including metamorphosed sandstones and argillites, but are kaolinitic and contain more detrital and authigenic chert, quartz, and potash feldspar. Local derivation of these clasts seems unlikely. By comparing geochemical ratios with those defined for continental margin sandstones, and well characterised New Zealand tectonic terranes, we interpret the majority of clasts in the lower Torehina Formation to have been derived from a dissected orogen, with mixtures of felsic and volcanogenic-derived sediment. The most likely sources are the Waipapa and Torlesse Terranes. The remaining 20–30% of the clasts in the lower Torehina Formation were originally friable, are coarse grained, and appear to be lithologically exotic relative to known metamorphosed sandstones in basement terrane sources on North Island. Some clasts contain coal laminae and particles, and all contain detrital kaolinite as lithic fragments and matrix. Such characteristics imply a non-marine to marginal-marine source containing sediment derived from strongly weathered granite or granodiorite. Mechanical fragility implies a likely proximal, easily erodible source. We propose that this group of clasts was derived from an Upper Cretaceous sedimentary cover, either part of a locally developed basin fill or part of a once regionally extensive cover on North Island. Either case defines a more widely distributed Cretaceous source than found today
Community detection‐based deep neural network architectures: A fully automated framework based on Likert‐scale data.
Deep neural networks (DNNs) have emerged as a state‐of‐the‐art tool in very different research fields due to its adaptive power to the decision space since they do not presuppose any linear relationship between data. Some of the main disadvantages of these trending models are that the choice of the network underlying architecture profoundly influences the performance of the model and that the architecture design requires prior knowledge of the field of study. The use of questionnaires is hugely extended in social/behavioral sciences. The main contribution of this work is to automate the process of a DNN architecture design by using an agglomerative hierarchical algorithm that mimics the conceptual structure of such surveys. Although the train had regression purposes, it is easily convertible to deal with classification tasks. Our proposed methodology will be tested with a database containing socio‐demographic data and the responses to five psychometric Likert scales related to the prediction of happiness. These scales have been already used to design a DNN architecture based on the subdimension of the scales. We show that our new network configurations outperform the previous existing DNN architectures
User Intent Prediction in Information-seeking Conversations
Conversational assistants are being progressively adopted by the general
population. However, they are not capable of handling complicated
information-seeking tasks that involve multiple turns of information exchange.
Due to the limited communication bandwidth in conversational search, it is
important for conversational assistants to accurately detect and predict user
intent in information-seeking conversations. In this paper, we investigate two
aspects of user intent prediction in an information-seeking setting. First, we
extract features based on the content, structural, and sentiment
characteristics of a given utterance, and use classic machine learning methods
to perform user intent prediction. We then conduct an in-depth feature
importance analysis to identify key features in this prediction task. We find
that structural features contribute most to the prediction performance. Given
this finding, we construct neural classifiers to incorporate context
information and achieve better performance without feature engineering. Our
findings can provide insights into the important factors and effective methods
of user intent prediction in information-seeking conversations.Comment: Accepted to CHIIR 201
- …