19 Hits in 0.73 sec

A Deep Learning Framework for COVID Outbreak Prediction [article]

Neeraj, Jimson Mathew, Ranjan Kumar Behera, Zenin Easa Panthakkalakath
2020 arXiv   pre-print
As, time in a vector representation can be easily added with many architectures. This vector representation is called Time2Vec.  ...  It helps in detection on crucial temporal information, resulting in a highly interpretable network. Additionally, we implement a learnable vector embedding for time.  ...  T t α n t . h n t      (13) Time2Vec We used Time2Vec [38] , a representation for the time which is invariant to time rescaling.  ... 
arXiv:2010.00382v2 fatcat:h2nanmlqonck3gbf6gqvapl5iu

Long-Range Transformers for Dynamic Spatiotemporal Forecasting [article]

Jake Grigsby, Zhe Wang, Yanjun Qi
2021 arXiv   pre-print
Multivariate Time Series Forecasting (TSF) focuses on the prediction of future values based on historical context.  ...  This paper addresses the problem by translating multivariate TSF into a novel spatiotemporal sequence formulation where each input token represents the value of a single variable at a given timestep.  ...  Time2Vec passes a representation of absolute time (e.g., the calendar datetime) through sinusoidal patterns of learned offsets and wavelengths.  ... 
arXiv:2109.12218v1 fatcat:c3djclmyhrf6nkpyh3vz4wrw4i

Encoder-Decoder Architecture for Supervised Dynamic Graph Learning: A Survey [article]

Yuecai Zhu, Fuyuan Lyu, Chengming Hu, Xi Chen, Xue Liu
2022 arXiv   pre-print
to interpret the learning of temporal information with a generalized framework.  ...  In recent years, the prevalent online services generate a sheer volume of user activity data.  ...  Time2Vec aims to generate a simple vector representation of time so as to enable different learning algorithms to learn the temporal correlation as well as the periodicity with the explicit use of time  ... 
arXiv:2203.10480v2 fatcat:tf7n73rhtbbcpptbn6lyvhcew4

Knowledge-enhanced Session-based Recommendation with Temporal Transformer [article]

Rongzhi Zhang, Yulong Gu, Xiaoyu Shen, Hui Su
2021 arXiv   pre-print
Specifically, a knowledge graph, which models contexts among items within a session and their corresponding attributes, is proposed to obtain item embeddings through graph representation learning.  ...  In this paper, we propose a framework called Knowledge-enhanced Session-based Recommendation with Temporal Transformer (KSTT) to incorporate such information when learning the item and session embeddings  ...  Mercer Time Embedding(MTE). Mercer Time Embedding(MTE) learns time representation with transition-invariant property [26] .  ... 
arXiv:2112.08745v1 fatcat:epftrmha7fdkvm5tg2bvotimna

Predicting Human Strategies in Simulated Search and Rescue Task [article]

Vidhi Jain, Rohit Jena, Huao Li, Tejus Gupta, Dana Hughes, Michael Lewis, Katia Sycara
2020 arXiv   pre-print
In a search and rescue scenario, rescuers may have different knowledge of the environment and strategies for exploration.  ...  As the neural networks are data-driven, we design a diverse set of artificial "faux human" agents for training, to test them with limited human rescuer trajectory data.  ...  Acknowledgments and Disclosure of Funding  ... 
arXiv:2011.07656v2 fatcat:e5xu6jpne5h3lfcln7wtd2vbgy

CRNNs for Urban Sound Tagging with spatiotemporal context [article]

Augustin Arnault, Nicolas Riche
2020 arXiv   pre-print
This paper describes CRNNs we used to participate in Task 5 of the DCASE 2020 challenge. This task focuses on hierarchical multilabel urban sound tagging with spatiotemporal context.  ...  Metadata embeddings In [8] , the author provides a model-agnostic vector representation for time, called Time2Vec, that can be easily imported into many existing and future architectures and improve their  ...  System1 outputs a prediction vector of 31 sources of noise (8 coarse-grained tags + 23 fine-grained tags).  ... 
arXiv:2008.10413v2 fatcat:shnkufbwlfet7j2q2v6wnxjoiu

Learning User Embeddings from Temporal Social Media Data: A Survey [article]

Fatema Hasan, Kevin S. Xu, James R. Foulds, Shimei Pan
2021 arXiv   pre-print
In this paper, we survey representative work on learning a concise latent user representation (a.k.a. user embedding) that can capture the main characteristics of a social media user.  ...  In this survey, we focus on research that bridges the gap by incorporating temporal/sequential information in user representation learning.  ...  Generalizable time embedding Currently, except for Time2Vec [Kazemi et al., 2019] , there has not been much work that produces a model-agnostic vector representation of time that can easily be incorporated  ... 
arXiv:2105.07996v1 fatcat:6elhasieuzce5ogddsl7uukv64

Training for the Future: A Simple Gradient Interpolation Loss to Generalize Along Time [article]

Anshul Nasery, Soumyadeep Thakur, Vihari Piratla, Abir De, Sunita Sarawagi
2021 arXiv   pre-print
of time-invariant features.  ...  In this context, there is much prior work on enhancing temporal generalization, e.g. continuous transportation of past data, kernel smoothed time-sensitive parameters and more recently, adversarial learning  ...  To tackle this challenge, we obtain a representation vector τ t ∈ R m of time t by building upon an existing time embedding model TIME2VEC proposed in [25] , which is computed as, τ t [a] = ω a t + b  ... 
arXiv:2108.06721v2 fatcat:aqx6sfyhwzaxfjamxoatfyjvfm

Network representation learning systematic review: Ancestors and current development state

Amina Amara, Mohamed Ali Hadj Taieb, Mohamed Ben Aouicha
2021 Machine Learning with Applications  
Thus, we introduce a brief history of representation learning and word representation learning ancestor of network embedding.  ...  As a new learning paradigm, network representation learning has been proposed to map a real-world information network into a low-dimensional space while preserving inherent properties of the network.  ...  "Time2Vec: Learning a Vector Representation of Time." arXiv preprint arXiv:1907.05321 (2019). Niu, Liqiang, et al.  ... 
doi:10.1016/j.mlwa.2021.100130 fatcat:axhg2gxkzfds3icebro6hlman4

GTEA: Representation Learning for Temporal Interaction Graphs via Edge Aggregation [article]

Yiming Li, Da Sun Handason Tam, Siyue Xie, Xiaxin Liu, Qiu Fang Ying, Wing Cheong Lau, Dah Ming Chiu, Shou Zhi Chen
2020 arXiv   pre-print
We consider the problem of representation learning for temporal interaction graphs where a network of entities with complex interactions over an extended period of time is modeled as a graph with a rich  ...  To fully capture and model the dynamics of the network, we propose GTEA, a framework of representation learning for temporal interaction graphs with per-edge time-based aggregation.  ...  In this work, in order to learn edge embedding in irregular and continuous time-space, we consider integrating state-of-the-art sequence models with the recent time representation learning method -Time2Vec  ... 
arXiv:2009.05266v2 fatcat:catsc43x5fdozd4yfe2yfjmh3q

Topic Detection and Tracking with Time-Aware Document Embeddings [article]

Hang Jiang, Doug Beeferman, Weiquan Mao, Deb Roy
2021 arXiv   pre-print
In our work, we design a neural method that fuses temporal and textual information into a single representation of news documents for event detection.  ...  The time at which a message is communicated is a vital piece of metadata in many real-world natural language processing tasks such as Topic Detection and Tracking (TDT).  ...  is a pre-trained time encoder based on Time2Vec (Kazemi et al. 2019 ) that transforms a date and time into a dense vector, while preserving the timespecific characteristics (progression, periodicity,  ... 
arXiv:2112.06166v1 fatcat:daxoxqny2fgulo7sxvhvmjlxxe

TIGGER: Scalable Generative Modelling for Temporal Interaction Graphs [article]

Shubham Gupta, Sahil Manchanda, Srikanta Bedathur, Sayan Ranu
2022 arXiv   pre-print
There has been a recent surge in learning generative models for graphs.  ...  First, existing generative models do not scale with either the time horizon or the number of nodes.  ...  Next, to learn vector representation of time t ∈ R, we use following TIME2VEC (Kazemi et al. 2019) transformation. f t (t)[r] = ω r • t + ζ r , if r = 0 sin (ω r • t + ζ r ), 1 ≤ r < d T (5) where ω  ... 
arXiv:2203.03564v2 fatcat:qyzlttui6jamxiqmi4xzsmydbu

PinnerFormer: Sequence Modeling for User Representation at Pinterest [article]

Nikil Pancha, Andrew Zhai, Jure Leskovec, Charles Rosenberg
2022 arXiv   pre-print
Here we introduce PinnerFormer, a user representation trained to predict a user's future long-term engagement using a sequential model of a user's recent actions.  ...  and engagement when comparing PinnerFormer against our previous user representation.  ...  For each of these time features, we follow the common practice of encoding time using sine and cosine transformations with various periods in a manner similar to Time2vec [11] , but with 𝑃 fixed periods  ... 
arXiv:2205.04507v1 fatcat:yotytbjk2zbvviaicvax375rdu

Foundations and modelling of dynamic networks using Dynamic Graph Neural Networks: A survey

Joakim Skardinga, Bogdan Gabrys, Katarzyna Musial
2021 IEEE Access  
Meanwhile, graph neural networks (GNNs) have gained a lot of attention in recent years for their ability to perform well on a range of network science tasks, such as link prediction and node classification  ...  Dynamic networks are used in a wide range of fields, including social network analysis, recommender systems and epidemiology.  ...  Dimensions of a GNN produced hidden feature vector k Dimensions of a RNN/self-attention produced hidden feature vector X t Feature matrix at time t A t Adjacency matrix at time t A Predicted adjacency  ... 
doi:10.1109/access.2021.3082932 fatcat:4pbp2kn6ovf65pnm5pbv7idpim

Expect: EXplainable Prediction Model for Energy ConsumpTion

Amira Mouakher, Wissem Inoubli, Chahinez Ounoughi, Andrea Ko
2022 Mathematics  
In this paper, we propose an explainable deep learning model, called Expect, to forecast energy consumption from time series effectively.  ...  With the steady growth of energy demands and resource depletion in today's world, energy prediction models have gained more and more attention recently.  ...  Embedding We adopt the Time2Vec proposed in [24] as our model's embedding representation of time and weather features.  ... 
doi:10.3390/math10020248 fatcat:5unlecalgfbpbhzja356p5ekoq
« Previous Showing results 1 — 15 out of 19 results