MM-CPred: A Multi-task Predictive Model for Continuous-Time Event Sequences with Mixture Learning Losses View Full Text


Ontology type: schema:Chapter     


Chapter Info

DATE

2021-04-06

AUTHORS

Li Lin , Zan Zong , Lijie Wen , Chen Qian , Shuang Li , Jianmin Wang

ABSTRACT

Sequence prediction is a well-defined problem with a proliferation of applications, such as recommendation systems, social media monitor, economic analysis, etc. Recently, RNN-based methodologies have shown their superiority in time-series data analysis and sequence modeling. The question of which event would happen next is not difficult to answer anymore, but the prediction of when it would happen is still a mountain to climb. In this paper, we propose a Multi-task model to predict both event and their continuous timestamps at the same time. Specifically, (1) we design a two-layer RNN encoder for event sequences and a CNN encoder for time sequences, both equipped with multi-head self-attention to align history features; (2) we form multiple generative adversarial models for predicting future time sequences to solve the problem of multi-modal time distribution; (3) Mixture learning losses are adopted to conduct a 3-step learning strategy for training our model, the cross-entropy loss for events, Huber loss and adversarial classification loss which induces the Wasserstein distance for times. Due to these characteristics, we name it MM-CPred. The experiments on 4 real-life datasets confirmed its improvements compared with the baselines. More... »

PAGES

509-525

Book

TITLE

Database Systems for Advanced Applications

ISBN

978-3-030-73193-9
978-3-030-73194-6

Identifiers

URI

http://scigraph.springernature.com/pub.10.1007/978-3-030-73194-6_34

DOI

http://dx.doi.org/10.1007/978-3-030-73194-6_34

DIMENSIONS

https://app.dimensions.ai/details/publication/pub.1136956148


Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
Incoming Citations Browse incoming citations for this publication using opencitations.net

JSON-LD is the canonical representation for SciGraph data.

TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

[
  {
    "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
    "about": [
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/01", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Mathematical Sciences", 
        "type": "DefinedTerm"
      }, 
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Information and Computing Sciences", 
        "type": "DefinedTerm"
      }, 
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0104", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Statistics", 
        "type": "DefinedTerm"
      }, 
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Artificial Intelligence and Image Processing", 
        "type": "DefinedTerm"
      }
    ], 
    "author": [
      {
        "affiliation": {
          "alternateName": "School of Software, Tsinghua University, Beijing, China", 
          "id": "http://www.grid.ac/institutes/grid.12527.33", 
          "name": [
            "School of Software, Tsinghua University, Beijing, China"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Lin", 
        "givenName": "Li", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "School of Software, Tsinghua University, Beijing, China", 
          "id": "http://www.grid.ac/institutes/grid.12527.33", 
          "name": [
            "School of Software, Tsinghua University, Beijing, China"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Zong", 
        "givenName": "Zan", 
        "id": "sg:person.011137150052.17", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011137150052.17"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "School of Software, Tsinghua University, Beijing, China", 
          "id": "http://www.grid.ac/institutes/grid.12527.33", 
          "name": [
            "School of Software, Tsinghua University, Beijing, China"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Wen", 
        "givenName": "Lijie", 
        "id": "sg:person.013640554311.55", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013640554311.55"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "School of Software, Tsinghua University, Beijing, China", 
          "id": "http://www.grid.ac/institutes/grid.12527.33", 
          "name": [
            "School of Software, Tsinghua University, Beijing, China"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Qian", 
        "givenName": "Chen", 
        "id": "sg:person.010475111706.64", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.010475111706.64"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "School of Software, Tsinghua University, Beijing, China", 
          "id": "http://www.grid.ac/institutes/grid.12527.33", 
          "name": [
            "School of Software, Tsinghua University, Beijing, China"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Li", 
        "givenName": "Shuang", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "School of Software, Tsinghua University, Beijing, China", 
          "id": "http://www.grid.ac/institutes/grid.12527.33", 
          "name": [
            "School of Software, Tsinghua University, Beijing, China"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Wang", 
        "givenName": "Jianmin", 
        "id": "sg:person.012303351315.43", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012303351315.43"
        ], 
        "type": "Person"
      }
    ], 
    "datePublished": "2021-04-06", 
    "datePublishedReg": "2021-04-06", 
    "description": "Sequence prediction is a well-defined problem with a proliferation of applications, such as recommendation systems, social media monitor, economic analysis, etc. Recently, RNN-based methodologies have shown their superiority in time-series data analysis and sequence modeling. The question of which event would happen next is not difficult to answer anymore, but the prediction of when it would happen is still a mountain to climb. In this paper, we propose a Multi-task model to predict both event and their continuous timestamps at the same time. Specifically, (1) we design a two-layer RNN encoder for event sequences and a CNN encoder for time sequences, both equipped with multi-head self-attention to align history features; (2) we form multiple generative adversarial models for predicting future time sequences to solve the problem of multi-modal time distribution; (3) Mixture learning losses are adopted to conduct a 3-step learning strategy for training our model, the cross-entropy loss for events, Huber loss and adversarial classification loss which induces the Wasserstein distance for times. Due to these characteristics, we name it MM-CPred. The experiments on 4 real-life datasets confirmed its improvements compared with the baselines.", 
    "editor": [
      {
        "familyName": "Jensen", 
        "givenName": "Christian S.", 
        "type": "Person"
      }, 
      {
        "familyName": "Lim", 
        "givenName": "Ee-Peng", 
        "type": "Person"
      }, 
      {
        "familyName": "Yang", 
        "givenName": "De-Nian", 
        "type": "Person"
      }, 
      {
        "familyName": "Lee", 
        "givenName": "Wang-Chien", 
        "type": "Person"
      }, 
      {
        "familyName": "Tseng", 
        "givenName": "Vincent S.", 
        "type": "Person"
      }, 
      {
        "familyName": "Kalogeraki", 
        "givenName": "Vana", 
        "type": "Person"
      }, 
      {
        "familyName": "Huang", 
        "givenName": "Jen-Wei", 
        "type": "Person"
      }, 
      {
        "familyName": "Shen", 
        "givenName": "Chih-Ya", 
        "type": "Person"
      }
    ], 
    "genre": "chapter", 
    "id": "sg:pub.10.1007/978-3-030-73194-6_34", 
    "inLanguage": "en", 
    "isAccessibleForFree": false, 
    "isPartOf": {
      "isbn": [
        "978-3-030-73193-9", 
        "978-3-030-73194-6"
      ], 
      "name": "Database Systems for Advanced Applications", 
      "type": "Book"
    }, 
    "keywords": [
      "event sequences", 
      "multi-task model", 
      "proliferation of applications", 
      "generative adversarial model", 
      "real-life datasets", 
      "cross-entropy loss", 
      "time series data analysis", 
      "learning loss", 
      "CNN encoder", 
      "classification loss", 
      "recommendation system", 
      "RNN encoder", 
      "sequence modeling", 
      "adversarial model", 
      "Media Monitor", 
      "sequence prediction", 
      "Huber loss", 
      "time sequence", 
      "encoder", 
      "learning strategies", 
      "Wasserstein distance", 
      "data analysis", 
      "predictive model", 
      "RNN", 
      "timestamps", 
      "same time", 
      "dataset", 
      "model", 
      "superiority", 
      "prediction", 
      "applications", 
      "monitor", 
      "methodology", 
      "modeling", 
      "system", 
      "features", 
      "sequence", 
      "time distribution", 
      "time", 
      "history features", 
      "experiments", 
      "improvement", 
      "strategies", 
      "distance", 
      "analysis", 
      "events", 
      "characteristics", 
      "questions", 
      "loss", 
      "distribution", 
      "baseline", 
      "economic analysis", 
      "proliferation", 
      "Mountains", 
      "problem", 
      "paper", 
      "social media monitor", 
      "continuous timestamps", 
      "two-layer RNN encoder", 
      "align history features", 
      "multiple generative adversarial models", 
      "future time sequences", 
      "multi-modal time distribution", 
      "Mixture learning losses", 
      "adversarial classification loss", 
      "MM-CPred", 
      "Multi-task Predictive Model", 
      "Continuous-Time Event Sequences"
    ], 
    "name": "MM-CPred: A Multi-task Predictive Model for Continuous-Time Event Sequences with Mixture Learning Losses", 
    "pagination": "509-525", 
    "productId": [
      {
        "name": "dimensions_id", 
        "type": "PropertyValue", 
        "value": [
          "pub.1136956148"
        ]
      }, 
      {
        "name": "doi", 
        "type": "PropertyValue", 
        "value": [
          "10.1007/978-3-030-73194-6_34"
        ]
      }
    ], 
    "publisher": {
      "name": "Springer Nature", 
      "type": "Organisation"
    }, 
    "sameAs": [
      "https://doi.org/10.1007/978-3-030-73194-6_34", 
      "https://app.dimensions.ai/details/publication/pub.1136956148"
    ], 
    "sdDataset": "chapters", 
    "sdDatePublished": "2022-01-01T19:24", 
    "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
    "sdPublisher": {
      "name": "Springer Nature - SN SciGraph project", 
      "type": "Organization"
    }, 
    "sdSource": "s3://com-springernature-scigraph/baseset/20220101/entities/gbq_results/chapter/chapter_433.jsonl", 
    "type": "Chapter", 
    "url": "https://doi.org/10.1007/978-3-030-73194-6_34"
  }
]
 

Download the RDF metadata as:  json-ld nt turtle xml License info

HOW TO GET THIS DATA PROGRAMMATICALLY:

JSON-LD is a popular format for linked data which is fully compatible with JSON.

curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-73194-6_34'

N-Triples is a line-based linked data format ideal for batch operations.

curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-73194-6_34'

Turtle is a human-readable linked data format.

curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-73194-6_34'

RDF/XML is a standard XML format for linked data.

curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-73194-6_34'


 

This table displays all metadata directly associated to this object as RDF triples.

204 TRIPLES      23 PREDICATES      95 URIs      86 LITERALS      7 BLANK NODES

Subject Predicate Object
1 sg:pub.10.1007/978-3-030-73194-6_34 schema:about anzsrc-for:01
2 anzsrc-for:0104
3 anzsrc-for:08
4 anzsrc-for:0801
5 schema:author Ne9f583977a6c4035bb30b12acb657d92
6 schema:datePublished 2021-04-06
7 schema:datePublishedReg 2021-04-06
8 schema:description Sequence prediction is a well-defined problem with a proliferation of applications, such as recommendation systems, social media monitor, economic analysis, etc. Recently, RNN-based methodologies have shown their superiority in time-series data analysis and sequence modeling. The question of which event would happen next is not difficult to answer anymore, but the prediction of when it would happen is still a mountain to climb. In this paper, we propose a Multi-task model to predict both event and their continuous timestamps at the same time. Specifically, (1) we design a two-layer RNN encoder for event sequences and a CNN encoder for time sequences, both equipped with multi-head self-attention to align history features; (2) we form multiple generative adversarial models for predicting future time sequences to solve the problem of multi-modal time distribution; (3) Mixture learning losses are adopted to conduct a 3-step learning strategy for training our model, the cross-entropy loss for events, Huber loss and adversarial classification loss which induces the Wasserstein distance for times. Due to these characteristics, we name it MM-CPred. The experiments on 4 real-life datasets confirmed its improvements compared with the baselines.
9 schema:editor Na638f19488fa4c468410896453e34b5b
10 schema:genre chapter
11 schema:inLanguage en
12 schema:isAccessibleForFree false
13 schema:isPartOf Nb0281d5a975f446eb9b2c73b89264e7e
14 schema:keywords CNN encoder
15 Continuous-Time Event Sequences
16 Huber loss
17 MM-CPred
18 Media Monitor
19 Mixture learning losses
20 Mountains
21 Multi-task Predictive Model
22 RNN
23 RNN encoder
24 Wasserstein distance
25 adversarial classification loss
26 adversarial model
27 align history features
28 analysis
29 applications
30 baseline
31 characteristics
32 classification loss
33 continuous timestamps
34 cross-entropy loss
35 data analysis
36 dataset
37 distance
38 distribution
39 economic analysis
40 encoder
41 event sequences
42 events
43 experiments
44 features
45 future time sequences
46 generative adversarial model
47 history features
48 improvement
49 learning loss
50 learning strategies
51 loss
52 methodology
53 model
54 modeling
55 monitor
56 multi-modal time distribution
57 multi-task model
58 multiple generative adversarial models
59 paper
60 prediction
61 predictive model
62 problem
63 proliferation
64 proliferation of applications
65 questions
66 real-life datasets
67 recommendation system
68 same time
69 sequence
70 sequence modeling
71 sequence prediction
72 social media monitor
73 strategies
74 superiority
75 system
76 time
77 time distribution
78 time sequence
79 time series data analysis
80 timestamps
81 two-layer RNN encoder
82 schema:name MM-CPred: A Multi-task Predictive Model for Continuous-Time Event Sequences with Mixture Learning Losses
83 schema:pagination 509-525
84 schema:productId N3b89fbcfbc4e4fe7b13b7cf36f08a319
85 Nfd9b678017ad471d9473cdc635b4d115
86 schema:publisher N624e75f5719244278dd41fef07e3172b
87 schema:sameAs https://app.dimensions.ai/details/publication/pub.1136956148
88 https://doi.org/10.1007/978-3-030-73194-6_34
89 schema:sdDatePublished 2022-01-01T19:24
90 schema:sdLicense https://scigraph.springernature.com/explorer/license/
91 schema:sdPublisher N51c96d2ae7ff4a61a8600690ca1cf548
92 schema:url https://doi.org/10.1007/978-3-030-73194-6_34
93 sgo:license sg:explorer/license/
94 sgo:sdDataset chapters
95 rdf:type schema:Chapter
96 N020fb3d3ed1348acabdb5e329da046ed schema:familyName Kalogeraki
97 schema:givenName Vana
98 rdf:type schema:Person
99 N12cab18267a0448a8a86cce7d149b9d5 rdf:first N020fb3d3ed1348acabdb5e329da046ed
100 rdf:rest N985784acbd664ffe9928d3e8b07db3f4
101 N2490993471674000b0e3f612241f1af1 rdf:first N2b60771aeb844c4abd66cc71cd8d6bdc
102 rdf:rest N4f00b94294bf44ea922d9c9165a9a24e
103 N28b00497571a4c9c8a7ec2fe63bb2327 schema:familyName Tseng
104 schema:givenName Vincent S.
105 rdf:type schema:Person
106 N2b60771aeb844c4abd66cc71cd8d6bdc schema:familyName Lee
107 schema:givenName Wang-Chien
108 rdf:type schema:Person
109 N2c2472f573e14e03858ce2cfb8e6e7c7 schema:affiliation grid-institutes:grid.12527.33
110 schema:familyName Lin
111 schema:givenName Li
112 rdf:type schema:Person
113 N3b89fbcfbc4e4fe7b13b7cf36f08a319 schema:name doi
114 schema:value 10.1007/978-3-030-73194-6_34
115 rdf:type schema:PropertyValue
116 N49e42abe9b74451ba3228ac93b79e5fe rdf:first sg:person.013640554311.55
117 rdf:rest N7eb0d77a7fde4b79b22e68a3eb19ac4f
118 N4bca8889328a46f19a5d168cc18d458c rdf:first N833f9d3b54ee4ebb9e065de41ac70460
119 rdf:rest rdf:nil
120 N4f00b94294bf44ea922d9c9165a9a24e rdf:first N28b00497571a4c9c8a7ec2fe63bb2327
121 rdf:rest N12cab18267a0448a8a86cce7d149b9d5
122 N51c96d2ae7ff4a61a8600690ca1cf548 schema:name Springer Nature - SN SciGraph project
123 rdf:type schema:Organization
124 N586f72896d4c46f69dc010d96b821090 rdf:first sg:person.012303351315.43
125 rdf:rest rdf:nil
126 N624e75f5719244278dd41fef07e3172b schema:name Springer Nature
127 rdf:type schema:Organisation
128 N79adb267dba54c63b33ccf3310081b03 schema:familyName Lim
129 schema:givenName Ee-Peng
130 rdf:type schema:Person
131 N7eb0d77a7fde4b79b22e68a3eb19ac4f rdf:first sg:person.010475111706.64
132 rdf:rest Na25ab5d65f8d4629940db55de7fc56eb
133 N833f9d3b54ee4ebb9e065de41ac70460 schema:familyName Shen
134 schema:givenName Chih-Ya
135 rdf:type schema:Person
136 N94e3d769a79647c1a68d48420563db4f schema:familyName Huang
137 schema:givenName Jen-Wei
138 rdf:type schema:Person
139 N985784acbd664ffe9928d3e8b07db3f4 rdf:first N94e3d769a79647c1a68d48420563db4f
140 rdf:rest N4bca8889328a46f19a5d168cc18d458c
141 Na25ab5d65f8d4629940db55de7fc56eb rdf:first Nff9ece95c34246288c6777ce100948ad
142 rdf:rest N586f72896d4c46f69dc010d96b821090
143 Na638f19488fa4c468410896453e34b5b rdf:first Nea0bd4e3e8484c4db857bd51c5b33e11
144 rdf:rest Nd8e34bac807549e6878e649ea2caf484
145 Nb0281d5a975f446eb9b2c73b89264e7e schema:isbn 978-3-030-73193-9
146 978-3-030-73194-6
147 schema:name Database Systems for Advanced Applications
148 rdf:type schema:Book
149 Nb0efc67e2ffc4e9198f03184a0a3b5e7 rdf:first sg:person.011137150052.17
150 rdf:rest N49e42abe9b74451ba3228ac93b79e5fe
151 Ncb980da707984180b742754e3888c2d0 schema:familyName Yang
152 schema:givenName De-Nian
153 rdf:type schema:Person
154 Nd8e34bac807549e6878e649ea2caf484 rdf:first N79adb267dba54c63b33ccf3310081b03
155 rdf:rest Ndc912983f75d49028afe96a78bcf9dc9
156 Ndc912983f75d49028afe96a78bcf9dc9 rdf:first Ncb980da707984180b742754e3888c2d0
157 rdf:rest N2490993471674000b0e3f612241f1af1
158 Ne9f583977a6c4035bb30b12acb657d92 rdf:first N2c2472f573e14e03858ce2cfb8e6e7c7
159 rdf:rest Nb0efc67e2ffc4e9198f03184a0a3b5e7
160 Nea0bd4e3e8484c4db857bd51c5b33e11 schema:familyName Jensen
161 schema:givenName Christian S.
162 rdf:type schema:Person
163 Nfd9b678017ad471d9473cdc635b4d115 schema:name dimensions_id
164 schema:value pub.1136956148
165 rdf:type schema:PropertyValue
166 Nff9ece95c34246288c6777ce100948ad schema:affiliation grid-institutes:grid.12527.33
167 schema:familyName Li
168 schema:givenName Shuang
169 rdf:type schema:Person
170 anzsrc-for:01 schema:inDefinedTermSet anzsrc-for:
171 schema:name Mathematical Sciences
172 rdf:type schema:DefinedTerm
173 anzsrc-for:0104 schema:inDefinedTermSet anzsrc-for:
174 schema:name Statistics
175 rdf:type schema:DefinedTerm
176 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
177 schema:name Information and Computing Sciences
178 rdf:type schema:DefinedTerm
179 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
180 schema:name Artificial Intelligence and Image Processing
181 rdf:type schema:DefinedTerm
182 sg:person.010475111706.64 schema:affiliation grid-institutes:grid.12527.33
183 schema:familyName Qian
184 schema:givenName Chen
185 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.010475111706.64
186 rdf:type schema:Person
187 sg:person.011137150052.17 schema:affiliation grid-institutes:grid.12527.33
188 schema:familyName Zong
189 schema:givenName Zan
190 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011137150052.17
191 rdf:type schema:Person
192 sg:person.012303351315.43 schema:affiliation grid-institutes:grid.12527.33
193 schema:familyName Wang
194 schema:givenName Jianmin
195 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012303351315.43
196 rdf:type schema:Person
197 sg:person.013640554311.55 schema:affiliation grid-institutes:grid.12527.33
198 schema:familyName Wen
199 schema:givenName Lijie
200 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013640554311.55
201 rdf:type schema:Person
202 grid-institutes:grid.12527.33 schema:alternateName School of Software, Tsinghua University, Beijing, China
203 schema:name School of Software, Tsinghua University, Beijing, China
204 rdf:type schema:Organization
 




Preview window. Press ESC to close (or click here)


...