RI: AF: Medium: Learning and Matrix Reconstruction with the Max-Norm and Related Factorization Norms View Homepage


Ontology type: schema:MonetaryGrant     


Grant Info

YEARS

2013-2019

FUNDING AMOUNT

915986 USD

ABSTRACT

Matrix learning is fundamental in many learning problems. These include problems that can be directly formulated as learning some unknown matrix, as well as a broader class of learning problems involving a matrix of parameters. The most direct matrix learning problem is matrix completion, completing unseen entries in a partially observed matrix. Matrix completion has recently received much attention both in practice in collaborative filtering (notably through the Netflix challenge), and theoretical analysis as an extension to compressed sensing. Matrix learning has also been used for clustering, transfer and multi-task learning, and similarity learning. The dominant approach to matrix learning in recent years, especially in the context of matrix completion, has used the matrix trace norm (developed in part by the PI on this award). Indeed, trace norm-based methods enjoy much success in a variety of applications. This project develops and studies alternative matrix norms to the trace-norm, most importantly the promising max-norm. Learning with the max-norm was initially presented in 2004 (along with the trace norm), but has not received the same attention, despite many theoretical and empirical advantages. This project identifies domains where the max-norm and related norms can be beneficial, develops computational methods for using these norms, and promotes the adoption of these norms. A central aim is to develop optimization methods for max-norm regularized problems that are as efficient as the corresponding methods for trace-norm regularized problems, such as singular value thresholding and LR-type methods. Beyond matrix completion, the project applies the max-norm both to problems where the trace-norm has been previously applied, and in novel settings. Novel applications include clustering, binary hashing, crowdsourcing, modeling rankings by a population, and similarity learning. Research under this project links the machine learning and theory-of-computation research communities (where SDP relaxations essentially corresponding to the max-norm have played a significant role in recent years). The project forms bridges between the communities, enabled in part by cross-disciplinary tutorials. Through collaboration with sociologists the PIs reach out to the social sciences, and increase the broad impact of the work by presenting it in an approachable and useable way to this audience. More... »

URL

http://www.nsf.gov/awardsearch/showAward?AWD_ID=1302662&HistoricalAwards=false

Related SciGraph Publications

  • 2017. An Improved Integrality Gap for the Călinescu-Karloff-Rabani Relaxation for Multiway Cut in INTEGER PROGRAMMING AND COMBINATORIAL OPTIMIZATION
  • 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/2008/2208", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "type": "DefinedTerm"
          }
        ], 
        "amount": {
          "currency": "USD", 
          "type": "MonetaryAmount", 
          "value": "915986"
        }, 
        "description": "Matrix learning is fundamental in many learning problems. These include problems that can be directly formulated as learning some unknown matrix, as well as a broader class of learning problems involving a matrix of parameters. The most direct matrix learning problem is matrix completion, completing unseen entries in a partially observed matrix. Matrix completion has recently received much attention both in practice in collaborative filtering (notably through the Netflix challenge), and theoretical analysis as an extension to compressed sensing. Matrix learning has also been used for clustering, transfer and multi-task learning, and similarity learning. The dominant approach to matrix learning in recent years, especially in the context of matrix completion, has used the matrix trace norm (developed in part by the PI on this award). Indeed, trace norm-based methods enjoy much success in a variety of applications. This project develops and studies alternative matrix norms to the trace-norm, most importantly the promising max-norm. Learning with the max-norm was initially presented in 2004 (along with the trace norm), but has not received the same attention, despite many theoretical and empirical advantages. This project identifies domains where the max-norm and related norms can be beneficial, develops computational methods for using these norms, and promotes the adoption of these norms. A central aim is to develop optimization methods for max-norm regularized problems that are as efficient as the corresponding methods for trace-norm regularized problems, such as singular value thresholding and LR-type methods. Beyond matrix completion, the project applies the max-norm both to problems where the trace-norm has been previously applied, and in novel settings. Novel applications include clustering, binary hashing, crowdsourcing, modeling rankings by a population, and similarity learning. Research under this project links the machine learning and theory-of-computation research communities (where SDP relaxations essentially corresponding to the max-norm have played a significant role in recent years). The project forms bridges between the communities, enabled in part by cross-disciplinary tutorials. Through collaboration with sociologists the PIs reach out to the social sciences, and increase the broad impact of the work by presenting it in an approachable and useable way to this audience.", 
        "endDate": "2019-03-31T00:00:00Z", 
        "funder": {
          "id": "https://www.grid.ac/institutes/grid.457785.c", 
          "type": "Organization"
        }, 
        "id": "sg:grant.3483837", 
        "identifier": [
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "3483837"
            ]
          }, 
          {
            "name": "nsf_id", 
            "type": "PropertyValue", 
            "value": [
              "1302662"
            ]
          }
        ], 
        "inLanguage": [
          "en"
        ], 
        "keywords": [
          "alternative matrix norms", 
          "METHODS", 
          "bridge", 
          "empirical advantages", 
          "award", 
          "computation research communities", 
          "multi-task learning", 
          "RI", 
          "same attention", 
          "practice", 
          "research", 
          "broad class", 
          "ranking", 
          "adoption", 
          "cross-disciplinary tutorials", 
          "direct matrix learning problem", 
          "learning", 
          "useable way", 
          "context", 
          "trace-norm", 
          "computational methods", 
          "AF", 
          "collaboration", 
          "transfer", 
          "trace norm", 
          "related norms", 
          "LR-type methods", 
          "crowdsourcing", 
          "machine learning", 
          "theory", 
          "part", 
          "matrix trace norm", 
          "project", 
          "clustering", 
          "community", 
          "much attention", 
          "novel settings", 
          "matrix completion", 
          "SDP relaxation", 
          "Pi", 
          "audience", 
          "theoretical analysis", 
          "much success", 
          "domain", 
          "unknown matrix", 
          "Netflix challenge", 
          "corresponding methods", 
          "extension", 
          "dominant approach", 
          "compressed sensing", 
          "parameters", 
          "singular value thresholding", 
          "max-norm", 
          "sociologists", 
          "central aim", 
          "social sciences", 
          "many learning problems", 
          "matrix", 
          "work", 
          "optimization method", 
          "matrix learning", 
          "population", 
          "similarity learning", 
          "Related Factorization Norms", 
          "novel application", 
          "collaborative filtering", 
          "application", 
          "medium", 
          "promising max-norm", 
          "unseen entries", 
          "problem", 
          "recent years", 
          "variety", 
          "broad impact", 
          "PIs", 
          "matrix reconstruction", 
          "binary hashing", 
          "norms", 
          "significant role"
        ], 
        "name": "RI: AF: Medium: Learning and Matrix Reconstruction with the Max-Norm and Related Factorization Norms", 
        "recipient": [
          {
            "id": "https://www.grid.ac/institutes/grid.287491.1", 
            "type": "Organization"
          }, 
          {
            "affiliation": {
              "id": "https://www.grid.ac/institutes/grid.287491.1", 
              "name": "Toyota Technological Institute at Chicago", 
              "type": "Organization"
            }, 
            "familyName": "Srebro", 
            "givenName": "Nathan", 
            "id": "sg:person.010235055202.41", 
            "type": "Person"
          }, 
          {
            "member": "sg:person.010235055202.41", 
            "roleName": "PI", 
            "type": "Role"
          }, 
          {
            "affiliation": {
              "id": "https://www.grid.ac/institutes/grid.287491.1", 
              "name": "Toyota Technological Institute at Chicago", 
              "type": "Organization"
            }, 
            "familyName": "Makarychev", 
            "givenName": "Yury", 
            "id": "sg:person.014753561577.44", 
            "type": "Person"
          }, 
          {
            "member": "sg:person.014753561577.44", 
            "roleName": "Co-PI", 
            "type": "Role"
          }
        ], 
        "sameAs": [
          "https://app.dimensions.ai/details/grant/grant.3483837"
        ], 
        "sdDataset": "grants", 
        "sdDatePublished": "2019-03-07T12:36", 
        "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/nsf_projects_6.xml.gz", 
        "startDate": "2013-06-01T00:00:00Z", 
        "type": "MonetaryGrant", 
        "url": "http://www.nsf.gov/awardsearch/showAward?AWD_ID=1302662&HistoricalAwards=false"
      }
    ]
     

    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.3483837'

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

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/grant.3483837'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/grant.3483837'

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

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/grant.3483837'


     

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

    132 TRIPLES      19 PREDICATES      103 URIs      94 LITERALS      6 BLANK NODES

    Subject Predicate Object
    1 sg:grant.3483837 schema:about anzsrc-for:2208
    2 schema:amount N5d6e82e02faa448cbf97c2ae4659d7d3
    3 schema:description Matrix learning is fundamental in many learning problems. These include problems that can be directly formulated as learning some unknown matrix, as well as a broader class of learning problems involving a matrix of parameters. The most direct matrix learning problem is matrix completion, completing unseen entries in a partially observed matrix. Matrix completion has recently received much attention both in practice in collaborative filtering (notably through the Netflix challenge), and theoretical analysis as an extension to compressed sensing. Matrix learning has also been used for clustering, transfer and multi-task learning, and similarity learning. The dominant approach to matrix learning in recent years, especially in the context of matrix completion, has used the matrix trace norm (developed in part by the PI on this award). Indeed, trace norm-based methods enjoy much success in a variety of applications. This project develops and studies alternative matrix norms to the trace-norm, most importantly the promising max-norm. Learning with the max-norm was initially presented in 2004 (along with the trace norm), but has not received the same attention, despite many theoretical and empirical advantages. This project identifies domains where the max-norm and related norms can be beneficial, develops computational methods for using these norms, and promotes the adoption of these norms. A central aim is to develop optimization methods for max-norm regularized problems that are as efficient as the corresponding methods for trace-norm regularized problems, such as singular value thresholding and LR-type methods. Beyond matrix completion, the project applies the max-norm both to problems where the trace-norm has been previously applied, and in novel settings. Novel applications include clustering, binary hashing, crowdsourcing, modeling rankings by a population, and similarity learning. Research under this project links the machine learning and theory-of-computation research communities (where SDP relaxations essentially corresponding to the max-norm have played a significant role in recent years). The project forms bridges between the communities, enabled in part by cross-disciplinary tutorials. Through collaboration with sociologists the PIs reach out to the social sciences, and increase the broad impact of the work by presenting it in an approachable and useable way to this audience.
    4 schema:endDate 2019-03-31T00:00:00Z
    5 schema:funder https://www.grid.ac/institutes/grid.457785.c
    6 schema:identifier N31e25d07ae3e49389a580b3ce9716258
    7 N9ffe66a1566b46edacffe49aea786f4a
    8 schema:inLanguage en
    9 schema:keywords AF
    10 LR-type methods
    11 METHODS
    12 Netflix challenge
    13 PIs
    14 Pi
    15 RI
    16 Related Factorization Norms
    17 SDP relaxation
    18 adoption
    19 alternative matrix norms
    20 application
    21 audience
    22 award
    23 binary hashing
    24 bridge
    25 broad class
    26 broad impact
    27 central aim
    28 clustering
    29 collaboration
    30 collaborative filtering
    31 community
    32 compressed sensing
    33 computation research communities
    34 computational methods
    35 context
    36 corresponding methods
    37 cross-disciplinary tutorials
    38 crowdsourcing
    39 direct matrix learning problem
    40 domain
    41 dominant approach
    42 empirical advantages
    43 extension
    44 learning
    45 machine learning
    46 many learning problems
    47 matrix
    48 matrix completion
    49 matrix learning
    50 matrix reconstruction
    51 matrix trace norm
    52 max-norm
    53 medium
    54 much attention
    55 much success
    56 multi-task learning
    57 norms
    58 novel application
    59 novel settings
    60 optimization method
    61 parameters
    62 part
    63 population
    64 practice
    65 problem
    66 project
    67 promising max-norm
    68 ranking
    69 recent years
    70 related norms
    71 research
    72 same attention
    73 significant role
    74 similarity learning
    75 singular value thresholding
    76 social sciences
    77 sociologists
    78 theoretical analysis
    79 theory
    80 trace norm
    81 trace-norm
    82 transfer
    83 unknown matrix
    84 unseen entries
    85 useable way
    86 variety
    87 work
    88 schema:name RI: AF: Medium: Learning and Matrix Reconstruction with the Max-Norm and Related Factorization Norms
    89 schema:recipient N7646c6fb42734a5c9a9b08280bb75138
    90 N8af8851caad943f4b8804344f10d4cd3
    91 sg:person.010235055202.41
    92 sg:person.014753561577.44
    93 https://www.grid.ac/institutes/grid.287491.1
    94 schema:sameAs https://app.dimensions.ai/details/grant/grant.3483837
    95 schema:sdDatePublished 2019-03-07T12:36
    96 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    97 schema:sdPublisher N955a7e44065a432d8eb0b1ceef75f426
    98 schema:startDate 2013-06-01T00:00:00Z
    99 schema:url http://www.nsf.gov/awardsearch/showAward?AWD_ID=1302662&HistoricalAwards=false
    100 sgo:license sg:explorer/license/
    101 sgo:sdDataset grants
    102 rdf:type schema:MonetaryGrant
    103 N31e25d07ae3e49389a580b3ce9716258 schema:name dimensions_id
    104 schema:value 3483837
    105 rdf:type schema:PropertyValue
    106 N5d6e82e02faa448cbf97c2ae4659d7d3 schema:currency USD
    107 schema:value 915986
    108 rdf:type schema:MonetaryAmount
    109 N7646c6fb42734a5c9a9b08280bb75138 schema:member sg:person.010235055202.41
    110 schema:roleName PI
    111 rdf:type schema:Role
    112 N8af8851caad943f4b8804344f10d4cd3 schema:member sg:person.014753561577.44
    113 schema:roleName Co-PI
    114 rdf:type schema:Role
    115 N955a7e44065a432d8eb0b1ceef75f426 schema:name Springer Nature - SN SciGraph project
    116 rdf:type schema:Organization
    117 N9ffe66a1566b46edacffe49aea786f4a schema:name nsf_id
    118 schema:value 1302662
    119 rdf:type schema:PropertyValue
    120 anzsrc-for:2208 schema:inDefinedTermSet anzsrc-for:
    121 rdf:type schema:DefinedTerm
    122 sg:person.010235055202.41 schema:affiliation https://www.grid.ac/institutes/grid.287491.1
    123 schema:familyName Srebro
    124 schema:givenName Nathan
    125 rdf:type schema:Person
    126 sg:person.014753561577.44 schema:affiliation https://www.grid.ac/institutes/grid.287491.1
    127 schema:familyName Makarychev
    128 schema:givenName Yury
    129 rdf:type schema:Person
    130 https://www.grid.ac/institutes/grid.287491.1 schema:name Toyota Technological Institute at Chicago
    131 rdf:type schema:Organization
    132 https://www.grid.ac/institutes/grid.457785.c schema:Organization
     




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


    ...