Essec\Faculty\Model\Contribution {#2216 ▼
#_index: "academ_contributions"
#_id: "10692"
#_source: array:26 [
"id" => "10692"
"slug" => "10692-variety-and-veracity-of-the-data-in-matrix-completion"
"yearMonth" => "2018-06"
"year" => "2018"
"title" => "Variety and Veracity of the Data in Matrix Completion"
"description" => "KLOPP, O., LOUNICI, K., TSYBAKOV, A. et ALAYA, M. (2018). Variety and Veracity of the Data in Matrix Completion. Dans: The 40th Conference on Stochastic Processes and their Applications. Gothenburg. 
KLOPP, O., LOUNICI, K., TSYBAKOV, A. et ALAYA, M. (2018). Variety and Veracity of the Data in Matrix "
"authors" => array:4 [
0 => array:3 [
"name" => "KLOPP Olga"
"bid" => "B00732676"
"slug" => "klopp-olga"
]
1 => array:1 [
"name" => "LOUNICI Karim"
]
2 => array:1 [
"name" => "TSYBAKOV Alexandre"
]
3 => array:1 [
"name" => "ALAYA Mokhtar"
]
]
"ouvrage" => "The 40th Conference on Stochastic Processes and their Applications"
"keywords" => array:4 [
0 => "high-dimensional prediction"
1 => "matrix completion"
2 => "low-rank matrix estimation"
3 => """
robust\n
estimation
"""
]
"updatedAt" => "2021-09-24 10:33:27"
"publicationUrl" => null
"publicationInfo" => array:3 [
"pages" => null
"volume" => null
"number" => null
]
"type" => array:2 [
"fr" => "Communications dans une conférence"
"en" => "Presentations at an Academic or Professional conference"
]
"support_type" => array:2 [
"fr" => null
"en" => null
]
"countries" => array:2 [
"fr" => null
"en" => null
]
"abstract" => array:2 [
"fr" => """
Beyond volume, variety and veracity are two important issues of the modern data. In this talk we discuss\n Beyond volume, variety and veracity are two important issues of the modern data. In this talk we dis
these questions in the context of the matrix completion problem. First, we considers the problem of estimation of\n these questions in the context of the matrix completion problem. First, we considers the problem of
a low-rank matrix when most of its entries are not observed and some of the observed entries are corrupted. The\n a low-rank matrix when most of its entries are not observed and some of the observed entries are cor
observations are noisy realizations of a sum of a low-rank matrix, which we wish to estimate, and a second matrix\n observations are noisy realizations of a sum of a low-rank matrix, which we wish to estimate, and a
having a complementary sparse structure such as elementwise sparsity or columnwise sparsity. We analyze a class of\n having a complementary sparse structure such as elementwise sparsity or columnwise sparsity. We anal
estimators obtained as solutions of a constrained convex optimization problem combining the nuclear norm penalty\n estimators obtained as solutions of a constrained convex optimization problem combining the nuclear
and a convex relaxation penalty for the sparse constraint.\n
In practical situations, data is often obtained from multiple sources which results in a collection of matrices\n In practical situations, data is often obtained from multiple sources which results in a collection
rather a single one. In the second part, we consider the problem of collective matrix completion with multiple and\n rather a single one. In the second part, we consider the problem of collective matrix completion wit
heterogeneous matrices, which can be count, binary, continuous, etc. We first investigate the setting where, for each\n heterogeneous matrices, which can be count, binary, continuous, etc. We first investigate the settin
source, the matrix entries are sampled from an exponential family distribution. Then we deal with the distribution-\n source, the matrix entries are sampled from an exponential family distribution. Then we deal with th
free setting. The estimation procedures are based on the penalized nuclear norm estimators. We prove that the\n free setting. The estimation procedures are based on the penalized nuclear norm estimators. We prove
proposed estimators achieve fast rates of convergence under the two considered setting.
"""
"en" => """
Beyond volume, variety and veracity are two important issues of the modern data. In this talk we discuss\n Beyond volume, variety and veracity are two important issues of the modern data. In this talk we dis
these questions in the context of the matrix completion problem. First, we considers the problem of estimation of\n these questions in the context of the matrix completion problem. First, we considers the problem of
a low-rank matrix when most of its entries are not observed and some of the observed entries are corrupted. The\n a low-rank matrix when most of its entries are not observed and some of the observed entries are cor
observations are noisy realizations of a sum of a low-rank matrix, which we wish to estimate, and a second matrix\n observations are noisy realizations of a sum of a low-rank matrix, which we wish to estimate, and a
having a complementary sparse structure such as elementwise sparsity or columnwise sparsity. We analyze a class of\n having a complementary sparse structure such as elementwise sparsity or columnwise sparsity. We anal
estimators obtained as solutions of a constrained convex optimization problem combining the nuclear norm penalty\n estimators obtained as solutions of a constrained convex optimization problem combining the nuclear
and a convex relaxation penalty for the sparse constraint.\n
In practical situations, data is often obtained from multiple sources which results in a collection of matrices\n In practical situations, data is often obtained from multiple sources which results in a collection
rather a single one. In the second part, we consider the problem of collective matrix completion with multiple and\n rather a single one. In the second part, we consider the problem of collective matrix completion wit
heterogeneous matrices, which can be count, binary, continuous, etc. We first investigate the setting where, for each\n heterogeneous matrices, which can be count, binary, continuous, etc. We first investigate the settin
source, the matrix entries are sampled from an exponential family distribution. Then we deal with the distribution-\n source, the matrix entries are sampled from an exponential family distribution. Then we deal with th
free setting. The estimation procedures are based on the penalized nuclear norm estimators. We prove that the\n free setting. The estimation procedures are based on the penalized nuclear norm estimators. We prove
proposed estimators achieve fast rates of convergence under the two considered setting.
"""
]
"authors_fields" => array:2 [
"fr" => "Systèmes d'Information, Data Analytics et Opérations"
"en" => "Information Systems, Data Analytics and Operations"
]
"indexedAt" => "2025-04-12T06:21:41.000Z"
"docTitle" => "Variety and Veracity of the Data in Matrix Completion"
"docSurtitle" => "Presentations at an Academic or Professional conference"
"authorNames" => "<a href="/cv/klopp-olga">KLOPP Olga</a>, LOUNICI Karim, TSYBAKOV Alexandre, ALAYA Mokhtar"
"docDescription" => "<span class="document-property-authors">KLOPP Olga, LOUNICI Karim, TSYBAKOV Alexandre, ALAYA Mokhtar</span><br><span class="document-property-authors_fields">Information Systems, Data Analytics and Operations</span> | <span class="document-property-year">2018</span> <span class="document-property-authors">KLOPP Olga, LOUNICI Karim, TSYBAKOV Alexandre, ALAYA Mokhtar "
"keywordList" => """
<a href="#">high-dimensional prediction</a>, <a href="#">matrix completion</a>, <a href="#">low-rank matrix estimation</a>, <a href="#">robust\n <a href="#">high-dimensional prediction</a>, <a href="#">matrix completion</a>, <a href="#">low-rank
estimation</a>
"""
"docPreview" => "<b>Variety and Veracity of the Data in Matrix Completion</b><br><span>2018-06 | Presentations at an Academic or Professional conference </span> <b>Variety and Veracity of the Data in Matrix Completion</b><br><span>2018-06 | Presentations at an "
"docType" => "research"
"publicationLink" => "<a href="#" target="_blank">Variety and Veracity of the Data in Matrix Completion</a>"
]
+lang: "en"
+"_type": "_doc"
+"_score": 8.971491
+"parent": null
}
