Research on Key Technology of Safe Machine Learning in Adversary Environment View Homepage


Ontology type: schema:MonetaryGrant     


Grant Info

YEARS

2012-2014

FUNDING AMOUNT

240000 CNY

ABSTRACT

Statistical machine learning can effectively find hidden patterns in empirical data, has been applied to many real world problems, so that the relevant application system with adaptive features. However, in the presence of malicious opponents of the enemy environment (such as intrusion detection, spam filtering, etc.), the introduction of machine learning will also bring new vulnerabilities to the relevant application system. A malicious opponent can exploit the nature of the machine to rely on the characteristics of the sample to attack, resulting in a decline in its classification performance, including: malicious opponents to explore the classification of the learner boundary, so that the learner will misclassify the sample; or malicious opponents pollution training samples, The classification accuracy of the classification decreased. In the face of this threat, this project studies some key technologies to enhance the safety of machine learning when there are malicious opponents. The main research contents include: (1) attack threat model and machine learning security analysis framework; (2) Attack detection technology; ③ off-line attack tolerance learning algorithm; ④ online case of tolerance tolerance learning algorithm. The theoretical analysis and algorithm realization of safety machine learning in this project can promote the theoretical understanding of machine learning safety and help solve the application obstacles of machine learning in adversary environment and show important prospects in the field of malicious behavior detection. More... »

URL

http://npd.nsfc.gov.cn/projectDetail.action?pid=61105050

Related SciGraph Publications

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/2008/2201", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "type": "DefinedTerm"
      }, 
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/2208", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "type": "DefinedTerm"
      }
    ], 
    "amount": {
      "currency": "CNY", 
      "type": "MonetaryAmount", 
      "value": "240000"
    }, 
    "description": "Statistical machine learning can effectively find hidden patterns in empirical data, has been applied to many real world problems, so that the relevant application system with adaptive features. However, in the presence of malicious opponents of the enemy environment (such as intrusion detection, spam filtering, etc.), the introduction of machine learning will also bring new vulnerabilities to the relevant application system. A malicious opponent can exploit the nature of the machine to rely on the characteristics of the sample to attack, resulting in a decline in its classification performance, including: malicious opponents to explore the classification of the learner boundary, so that the learner will misclassify the sample; or malicious opponents pollution training samples, The classification accuracy of the classification decreased. In the face of this threat, this project studies some key technologies to enhance the safety of machine learning when there are malicious opponents. The main research contents include: (1) attack threat model and machine learning security analysis framework; (2) Attack detection technology; \u2462 off-line attack tolerance learning algorithm; \u2463 online case of tolerance tolerance learning algorithm. The theoretical analysis and algorithm realization of safety machine learning in this project can promote the theoretical understanding of machine learning safety and help solve the application obstacles of machine learning in adversary environment and show important prospects in the field of malicious behavior detection.", 
    "endDate": "2014-12-31T00:00:00Z", 
    "funder": {
      "id": "https://www.grid.ac/institutes/grid.419696.5", 
      "type": "Organization"
    }, 
    "id": "sg:grant.7009604", 
    "identifier": [
      {
        "name": "dimensions_id", 
        "type": "PropertyValue", 
        "value": [
          "7009604"
        ]
      }, 
      {
        "name": "nsfc_id", 
        "type": "PropertyValue", 
        "value": [
          "61105050"
        ]
      }
    ], 
    "inLanguage": [
      "zh"
    ], 
    "keywords": [
      "line attack tolerance", 
      "classification accuracy", 
      "key technologies", 
      "machine learning security analysis framework", 
      "new vulnerabilities", 
      "hidden patterns", 
      "online case", 
      "research", 
      "main research contents", 
      "introduction", 
      "many real world problems", 
      "nature", 
      "samples", 
      "malicious opponents", 
      "theoretical understanding", 
      "safety", 
      "Safe Machine Learning", 
      "important prospects", 
      "empirical data", 
      "statistical machine learning", 
      "face", 
      "learners", 
      "detection technology", 
      "help", 
      "enemy environment", 
      "classification", 
      "adversary environment", 
      "learner boundary", 
      "presence", 
      "threat", 
      "relevant application system", 
      "algorithm", 
      "field", 
      "classification performance", 
      "machine", 
      "threat model", 
      "application obstacles", 
      "algorithm realization", 
      "decline", 
      "malicious behavior detection", 
      "tolerance tolerance", 
      "intrusion detection", 
      "machine learning", 
      "malicious opponents pollution training samples", 
      "spam filtering", 
      "characteristics", 
      "safety machine", 
      "theoretical analysis", 
      "adaptive features", 
      "project"
    ], 
    "name": "Research on Key Technology of Safe Machine Learning in Adversary Environment", 
    "recipient": [
      {
        "id": "https://www.grid.ac/institutes/grid.412110.7", 
        "type": "Organization"
      }, 
      {
        "affiliation": {
          "id": "https://www.grid.ac/institutes/grid.412110.7", 
          "name": "National University of Defense Technology", 
          "type": "Organization"
        }, 
        "familyName": "Long", 
        "givenName": "Jun", 
        "id": "sg:person.016662261561.62", 
        "type": "Person"
      }, 
      {
        "member": "sg:person.016662261561.62", 
        "roleName": "PI", 
        "type": "Role"
      }
    ], 
    "sameAs": [
      "https://app.dimensions.ai/details/grant/grant.7009604"
    ], 
    "sdDataset": "grants", 
    "sdDatePublished": "2019-03-07T12:39", 
    "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
    "sdPublisher": {
      "name": "Springer Nature - SN SciGraph project", 
      "type": "Organization"
    }, 
    "sdSource": "s3://com.uberresearch.data.processor/core_data/20181219_192338/projects/base/nsfc_projects_0.xml.gz", 
    "startDate": "2012-01-01T00:00:00Z", 
    "type": "MonetaryGrant", 
    "url": "http://npd.nsfc.gov.cn/projectDetail.action?pid=61105050"
  }
]
 

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/grant.7009604'

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

curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/grant.7009604'

Turtle is a human-readable linked data format.

curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/grant.7009604'

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

curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/grant.7009604'


 

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

97 TRIPLES      19 PREDICATES      73 URIs      64 LITERALS      5 BLANK NODES

Subject Predicate Object
1 sg:grant.7009604 schema:about anzsrc-for:2201
2 anzsrc-for:2208
3 schema:amount Naa0729de71a848938538fe25a5b47050
4 schema:description Statistical machine learning can effectively find hidden patterns in empirical data, has been applied to many real world problems, so that the relevant application system with adaptive features. However, in the presence of malicious opponents of the enemy environment (such as intrusion detection, spam filtering, etc.), the introduction of machine learning will also bring new vulnerabilities to the relevant application system. A malicious opponent can exploit the nature of the machine to rely on the characteristics of the sample to attack, resulting in a decline in its classification performance, including: malicious opponents to explore the classification of the learner boundary, so that the learner will misclassify the sample; or malicious opponents pollution training samples, The classification accuracy of the classification decreased. In the face of this threat, this project studies some key technologies to enhance the safety of machine learning when there are malicious opponents. The main research contents include: (1) attack threat model and machine learning security analysis framework; (2) Attack detection technology; ③ off-line attack tolerance learning algorithm; ④ online case of tolerance tolerance learning algorithm. The theoretical analysis and algorithm realization of safety machine learning in this project can promote the theoretical understanding of machine learning safety and help solve the application obstacles of machine learning in adversary environment and show important prospects in the field of malicious behavior detection.
5 schema:endDate 2014-12-31T00:00:00Z
6 schema:funder https://www.grid.ac/institutes/grid.419696.5
7 schema:identifier N4c78f6a032dd45c7b392c780856ee10e
8 Naa86baa39b0a420abed1a451cf31b9c9
9 schema:inLanguage zh
10 schema:keywords Safe Machine Learning
11 adaptive features
12 adversary environment
13 algorithm
14 algorithm realization
15 application obstacles
16 characteristics
17 classification
18 classification accuracy
19 classification performance
20 decline
21 detection technology
22 empirical data
23 enemy environment
24 face
25 field
26 help
27 hidden patterns
28 important prospects
29 introduction
30 intrusion detection
31 key technologies
32 learner boundary
33 learners
34 line attack tolerance
35 machine
36 machine learning
37 machine learning security analysis framework
38 main research contents
39 malicious behavior detection
40 malicious opponents
41 malicious opponents pollution training samples
42 many real world problems
43 nature
44 new vulnerabilities
45 online case
46 presence
47 project
48 relevant application system
49 research
50 safety
51 safety machine
52 samples
53 spam filtering
54 statistical machine learning
55 theoretical analysis
56 theoretical understanding
57 threat
58 threat model
59 tolerance tolerance
60 schema:name Research on Key Technology of Safe Machine Learning in Adversary Environment
61 schema:recipient N8c39207b3cb142749b2cf970325d8482
62 sg:person.016662261561.62
63 https://www.grid.ac/institutes/grid.412110.7
64 schema:sameAs https://app.dimensions.ai/details/grant/grant.7009604
65 schema:sdDatePublished 2019-03-07T12:39
66 schema:sdLicense https://scigraph.springernature.com/explorer/license/
67 schema:sdPublisher N983a337369684016a54813cbd9b2573b
68 schema:startDate 2012-01-01T00:00:00Z
69 schema:url http://npd.nsfc.gov.cn/projectDetail.action?pid=61105050
70 sgo:license sg:explorer/license/
71 sgo:sdDataset grants
72 rdf:type schema:MonetaryGrant
73 N4c78f6a032dd45c7b392c780856ee10e schema:name dimensions_id
74 schema:value 7009604
75 rdf:type schema:PropertyValue
76 N8c39207b3cb142749b2cf970325d8482 schema:member sg:person.016662261561.62
77 schema:roleName PI
78 rdf:type schema:Role
79 N983a337369684016a54813cbd9b2573b schema:name Springer Nature - SN SciGraph project
80 rdf:type schema:Organization
81 Naa0729de71a848938538fe25a5b47050 schema:currency CNY
82 schema:value 240000
83 rdf:type schema:MonetaryAmount
84 Naa86baa39b0a420abed1a451cf31b9c9 schema:name nsfc_id
85 schema:value 61105050
86 rdf:type schema:PropertyValue
87 anzsrc-for:2201 schema:inDefinedTermSet anzsrc-for:
88 rdf:type schema:DefinedTerm
89 anzsrc-for:2208 schema:inDefinedTermSet anzsrc-for:
90 rdf:type schema:DefinedTerm
91 sg:person.016662261561.62 schema:affiliation https://www.grid.ac/institutes/grid.412110.7
92 schema:familyName Long
93 schema:givenName Jun
94 rdf:type schema:Person
95 https://www.grid.ac/institutes/grid.412110.7 schema:name National University of Defense Technology
96 rdf:type schema:Organization
97 https://www.grid.ac/institutes/grid.419696.5 schema:Organization
 




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


...