What to Do Next: Modeling User Behaviors by Time-LSTM

By Y. Zhu, Hao Li, Yikang Liao, Beidou Wang, Ziyu Guan, Haifeng Liu, Deng Cai. International Joint Conference on Artificial Intelligence, 2017

499 citations recorded by the source corpus.

Recently, Recurrent Neural Network (RNN) solutions for recommender systems (RS) are becoming increasingly popular. The insight is that, there exist some intrinsic patterns in the sequence of users' actions, and RNN has been proved to perform excellently when modeling sequential data. In traditional tasks such as language modeling, RNN solutions usually only consider the sequential order of objects without the notion of interval. However, in RS, time intervals between users' actions are of significant importance in capturing the relations of users' actions and the traditional RNN architectures are not good at modeling them. In this paper, we propose a new LSTM variant, i.e. Time-LSTM, to model users' sequential actions. Time-LSTM equips LSTM with time gates to model time intervals. These time gates are specifically designed, so that compared to the traditional RNN solutions, Time-LSTM better captures both of users' short-term and long-term interests, so as to improve the recommendation performance. Experimental results on two real-world datasets show the superiority of the recommendation method using Time-LSTM over the traditional methods.

Subfields: Artificial Intelligence, Information Systems

View Paper: https://www.semanticscholar.org/paper/e1b93001202afd674cc3542cdb5599bff0d10157

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

Citations in the corpus

Where to Go Next: Modeling Long- and Short-Term User Preferences for Point-of-Interest Recommendation (2020, unknown): “Time-LSTM: (Zhu et al. 2017) This is an extension of LSTM which employs time gates to model the time intervals between continuous inputs.”
Relational Representation Learning for Dynamic (Knowledge) Graphs: A Survey (2019, background): “…problems [160, 9, 94, 158, 103, 200], where the duration between any two consecutive items in the sequence is considered equal, and several asynchronous sequence modeling problems [45, 143, 64, 167, 260, 100], RNNs have been the most common choice for extending sequence models to DTDGs and CTDGs.”
Self-attention with Functional Time Representation Learning (2019, methodology): “To deal with continuous time input in RNNs, a time-LSTM model was proposed with modified gate structures [27].”
Latent Cross: Making Use of Context in Recurrent Recommender Systems (2018, background): “Cross RNN [22, 34] [29, 32, 48] [17, 37, 46] [21, 43] [11] [20] [19, 33] [41] [49]”
Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks (2017, background): “Deep neural networks have been studied for time series forecasting [8, 27, 34, 37], i.”
Time2Vec: Learning a Vector Representation of Time (2019, methodology): “FM, we used the pre-processing used by Zhu et al. [61] to select ∼ 1 .”
STAMP: Short-Term Attention/Memory Priority Model for Session-based Recommendation (2018, background): “Another recent related work is the Time-LSTM model[21] which is a variant of the LSTM.”
Where to Go Next: A Spatio-Temporal Gated Network for Next POI Recommendation (2019, background): “Zhu et al.(2017) proposed a variant of Long-Short Term Memory network (LSTM), called Time-LSTM, to equip LSTM with time gates to model time intervals for next item recommendation.”
Predicting Dynamic Embedding Trajectory in Temporal Interaction Networks (2019, methodology): “same as item i’s embedding after its previous interaction with any user. Notice that this design decision is in stark contrast with the popular use of items’ one-hot vectors to update user embeddings [8, 45, 52], which has the following two disadvantages: (a) one-hot vector only contains the informa”
Dynamic Explainable Recommendation Based on Neural Attentive Models (2019, methodology): “…without considering the time interval information between two successive steps, which is yet an important signal for user behavior modeling, e.g., a user tend to have similar preferences within a short time, while large time gap may have more opportunities to change user interest (Zhu et al. 2017).”
Temporal Graph Networks for Deep Learning on Dynamic Graphs (2020, methodology): “In the latter case, the embeddings are either aggregated by taking a weighted sum [67, 74], ﬁt to time series models [30, 24, 11, 44], used as components in RNNs [55, 45, 41, 69, 8, 54, 48], or learned by imposing a smoothness constraint over time [33, 25, 67, 75, 73, 57, 22, 17, 50].”
STAN: Spatio-Temporal Attention Network for Next Location Recommendation (2021, methodology): “Then, Time-LSTM[45] proposes to add time gates to the LSTM structure to better adapt the spatiotemporal effect.”
Graph Neural Networks in Recommender Systems: A Survey (2020, background): “For example, Time-LSTM [83] extends the original LSTM structure with time gates and it greatly outperforms the original LSTM.”
Replicating a Trading Strategy by Means of LSTM for Financial Industry Applications (2018, unknown): “, see [19]–[21].”