Cold-start playlist recommendation with multitask learning

10.7287/peerj.preprints.27383v2 ◽

2018 ◽

Author(s):

Dawei Chen ◽

Cheng Soon Ong ◽

Aditya Krishna Menon

Keyword(s):

Binary Classification ◽

Optimal Solution ◽

Cold Start ◽

Multitask Learning ◽

Learning Objective ◽

Learning Method ◽

Empirical Results ◽

Bipartite Ranking ◽

Ranking Loss

Playlist recommendation involves producing a set of songs that a user might enjoy. We investigate this problem in three cold-start scenarios: (i) cold playlists, where we recommend songs to form new personalised playlists for an existing user; (ii) cold users, where we recommend songs to form new playlists for a new user; and (iii) cold songs, where we recommend newly released songs to extend users’ existing playlists. We propose a flexible multitask learning method to deal with all three settings. The method learns from user-curated playlists, and encourages songs in a playlist to be ranked higher than those that are not by minimising a bipartite ranking loss. Inspired by an equivalence between bipartite ranking and binary classification, we show how one can efficiently approximate an optimal solution of the multitask learning objective by minimising a classification loss. Empirical results on two real playlist datasets show the proposed approach has good performance for cold-start playlist recommendation.

Download Full-text

Cold-start playlist recommendation with multitask learning

10.7287/peerj.preprints.27383v1 ◽

2018 ◽

Author(s):

Dawei Chen ◽

Cheng Soon Ong ◽

Aditya Krishna Menon

Keyword(s):

Binary Classification ◽

Optimal Solution ◽

Cold Start ◽

Multitask Learning ◽

Learning Objective ◽

Learning Method ◽

Empirical Results ◽

Bipartite Ranking ◽

Ranking Loss

Playlist recommendation involves producing a set of songs that a user might enjoy. We investigate this problem in three cold-start scenarios: (i) cold playlists, where we recommend songs to form new personalised playlists for an existing user; (ii) cold users, where we recommend songs to form new playlists for a new user; and (iii) cold songs, where we recommend newly released songs to extend users’ existing playlists. We propose a flexible multitask learning method to deal with all three settings. The method learns from user-curated playlists, and encourages songs in a playlist to be ranked higher than those that are not by minimising a bipartite ranking loss. Inspired by an equivalence between bipartite ranking and binary classification, we show how one can efficiently approximate an optimal solution of the multitask learning objective by minimising a classification loss. Empirical results on two real playlist datasets show the proposed approach has good performance for cold-start playlist recommendation.

Download Full-text

Identifying Genetic related Neuroimaging Biomarkers of Alzheimer's Disease via Diagnosis-Guided Group Sparse Multitask Learning Method

Current Bioinformatics ◽

10.2174/1574893616666210708145303 ◽

2021 ◽

Vol 16 ◽

Author(s):

Anshi Lin ◽

Wei Kong ◽

Shuaiqun Wang

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Regression Model ◽

Multivariate Regression ◽

Prior Information ◽

Diagnostic Information ◽

Multitask Learning ◽

Learning Method ◽

Pathological Characteristics ◽

Different Types

Background: Advances in brain imaging and high-throughput genotyping techniques have provided new methods for studying the effects of genetic variation on brain structure and function. Traditionally, a variety of prior information has been added into the multivariate regression method for single nucleotide polymorphisms (SNPs) and quantitative traits (QTs) to improve the accuracy of prediction. In previous studies, brain regions of interest (ROIs) with different types of pathological characteristics (Alzheimer's Disease/Mild Cognitive Impairment/healthy control) can only be randomly dispersed in test cases, greatly limiting the prediction ability of the regression model and failing to obtain optimal global results. Objective: This study proposes a multivariate regression model informed by prior diagnostic information to overcome this limitation. Method: In the prediction model, we first consider traditional prior information and then design a new regularization form to integrate the diagnostic information of different sample ROIs into the model. Results: Experiments demonstrated that this method greatly improves the prediction accuracy of the model compared to other methods and selects a batch of promising pathogenic SNP loci. Conclusion: Taking into account that ROIs with different types of pathological characteristics can be employed as prior information, we propose a new method (Diagnosis-Guided Group Sparse Multitask Learning Method) that improves the ability to predict disease-related quantitative feature sites and select genetic feature factors, applying this model to research on the pathogenesis of Alzheimer's disease.

Download Full-text

Attribute Noise Robust Binary Classification (Student Abstract)

Proceedings of the AAAI Conference on Artificial Intelligence ◽

10.1609/aaai.v34i10.7221 ◽

2020 ◽

Vol 34 (10) ◽

pp. 13897-13898

Author(s):

Aditya Petety ◽

Sandhya Tripathi ◽

N. Hemachandra

Keyword(s):

Binary Classification ◽

Feature Space ◽

Noise Model ◽

Linear Classifiers ◽

Empirical Results ◽

Unknown Probability ◽

Attribute Noise ◽

Noise Robust ◽

Computational Intractability

We consider the problem of learning linear classifiers when both features and labels are binary. In addition, the features are noisy, i.e., they could be flipped with an unknown probability. In Sy-De attribute noise model, where all features could be noisy together with same probability, we show that 0-1 loss (l0−1) need not be robust but a popular surrogate, squared loss (lsq) is. In Asy-In attribute noise model, we prove that l0−1 is robust for any distribution over 2 dimensional feature space. However, due to computational intractability of l0−1, we resort to lsq and observe that it need not be Asy-In noise robust. Our empirical results support Sy-De robustness of squared loss for low to moderate noise rates.

Download Full-text

Multiple similarly effective solutions exist for biomedical feature selection and classification problems

Scientific Reports ◽

10.1038/s41598-017-13184-8 ◽

2017 ◽

Vol 7 (1) ◽

Cited By ~ 9

Author(s):

Jiamei Liu ◽

Cheng Xu ◽

Weifeng Yang ◽

Yayun Shu ◽

Weiwei Zheng ◽

...

Keyword(s):

Machine Learning ◽

Feature Selection ◽

Association Studies ◽

Binary Classification ◽

Learning Algorithms ◽

Optimal Solution ◽

Machine Learning Algorithms ◽

Disease Classification ◽

Genome Wide Association Studies ◽

Classification Problems

Abstract Binary classification is a widely employed problem to facilitate the decisions on various biomedical big data questions, such as clinical drug trials between treated participants and controls, and genome-wide association studies (GWASs) between participants with or without a phenotype. A machine learning model is trained for this purpose by optimizing the power of discriminating samples from two groups. However, most of the classification algorithms tend to generate one locally optimal solution according to the input dataset and the mathematical presumptions of the dataset. Here we demonstrated from the aspects of both disease classification and feature selection that multiple different solutions may have similar classification performances. So the existing machine learning algorithms may have ignored a horde of fishes by catching only a good one. Since most of the existing machine learning algorithms generate a solution by optimizing a mathematical goal, it may be essential for understanding the biological mechanisms for the investigated classification question, by considering both the generated solution and the ignored ones.

Download Full-text

Clustering by Errors: A Self-Organized Multitask Learning Method for Acoustic Scene Classification

Sensors ◽

10.3390/s22010036 ◽

2021 ◽

Vol 22 (1) ◽

pp. 36

Author(s):

Weiping Zheng ◽

Zhenyao Mo ◽

Gansen Zhao

Keyword(s):

Multitask Learning ◽

Classification Error ◽

Learning Method ◽

Scene Classification ◽

Fine Grained ◽

Learning Framework ◽

Self Organized ◽

Similarity Relations ◽

Hierarchy Construction ◽

Better Than

Acoustic scene classification (ASC) tries to inference information about the environment using audio segments. The inter-class similarity is a significant issue in ASC as acoustic scenes with different labels may sound quite similar. In this paper, the similarity relations amongst scenes are correlated with the classification error. A class hierarchy construction method by using classification error is then proposed and integrated into a multitask learning framework. The experiments have shown that the proposed multitask learning method improves the performance of ASC. On the TUT Acoustic Scene 2017 dataset, we obtain the ensemble fine-grained accuracy of 81.4%, which is better than the state-of-the-art. By using multitask learning, the basic Convolutional Neural Network (CNN) model can be improved by about 2.0 to 3.5 percent according to different spectrograms. The coarse category accuracies (for two to six super-classes) range from 77.0% to 96.2% by single models. On the revised version of the LITIS Rouen dataset, we achieve the ensemble fine-grained accuracy of 83.9%. The multitask learning models obtain an improvement of 1.6% to 1.8% compared to their basic models. The coarse category accuracies range from 94.9% to 97.9% for two to six super-classes with single models.

Download Full-text

Exploiting Past Users’ Interests and Predictions in an Active Learning Method for Dealing with Cold Start in Recommender Systems

Informatics ◽

10.3390/informatics5030035 ◽

2018 ◽

Vol 5 (3) ◽

pp. 35 ◽

Cited By ~ 2

Author(s):

Manuel Pozo ◽

Raja Chiky ◽

Farid Meziane ◽

Elisabeth Métais

Keyword(s):

Active Learning ◽

Recommender Systems ◽

Cold Start ◽

Learning Method ◽

Learning Techniques ◽

Learning Technique ◽

Active Learning Method ◽

Public Dataset

This paper focuses on the new users cold-start issue in the context of recommender systems. New users who do not receive pertinent recommendations may abandon the system. In order to cope with this issue, we use active learning techniques. These methods engage the new users to interact with the system by presenting them with a questionnaire that aims to understand their preferences to the related items. In this paper, we propose an active learning technique that exploits past users’ interests and past users’ predictions in order to identify the best questions to ask. Our technique achieves a better performance in terms of precision (RMSE), which leads to learn the users’ preferences in less questions. The experimentations were carried out in a small and public dataset to prove the applicability for handling cold start issues.

Download Full-text

CoupledCF: Learning Explicit and Implicit User-item Couplings in Recommendation for Deep Collaborative Filtering

Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2018/509 ◽

2018 ◽

Cited By ~ 9

Author(s):

Quangui Zhang ◽

Longbing Cao ◽

Chengzhang Zhu ◽

Zhiqiang Li ◽

Jinguang Sun

Keyword(s):

Collaborative Filtering ◽

Performance Improvement ◽

Recommender System ◽

Real World ◽

Matrix Factorization ◽

Cold Start ◽

Large Datasets ◽

Empirical Results ◽

The Rich ◽

Item Attributes

Non-IID recommender system discloses the nature of recommendation and has shown its potential in improving recommendation quality and addressing issues such as sparsity and cold start. It leverages existing work that usually treats users/items as in- dependent while ignoring the rich couplings within and between users and items, leading to limited performance improvement. In reality, users/items are related with various couplings existing within and between users and items, which may better ex- plain how and why a user has personalized pref- erence on an item. This work builds on non- IID learning to propose a neural user-item cou- pling learning for collaborative filtering, called CoupledCF. CoupledCF jointly learns explicit and implicit couplings within/between users and items w.r.t. user/item attributes and deep features for deep CF recommendation. Empirical results on two real-world large datasets show that CoupledCF significantly outperforms two latest neural recom- menders: neural matrix factorization and Google’s Wide&Deep network.

Download Full-text

Positive-Unlabeled Learning from Imbalanced Data

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2021/412 ◽

2021 ◽

Author(s):

Guangxin Su ◽

Weitong Chen ◽

Miao Xu

Keyword(s):

State Of The Art ◽

Binary Classification ◽

Estimation Error ◽

Imbalanced Data ◽

Classification Problem ◽

Learning Objective ◽

Financial Fraud ◽

Pu Learning ◽

Objective State ◽

Binary Classification Problem

Positive-unlabeled (PU) learning deals with the binary classification problem when only positive (P) and unlabeled (U) data are available, without negative (N) data. Existing PU methods perform well on the balanced dataset. However, in real applications such as financial fraud detection or medical diagnosis, data are always imbalanced. It remains unclear whether existing PU methods can perform well on imbalanced data. In this paper, we explore this problem and propose a general learning objective for PU learning targeting specially at imbalanced data. By this general learning objective, state-of-the-art PU methods based on optimizing a consistent risk can be adapted to conquer the imbalance. We theoretically show that in expectation, optimizing our learning objective is equivalent to learning a classifier on the oversampled balanced data with both P and N data available, and further provide an estimation error bound. Finally, experimental results validate the effectiveness of our proposal compared to state-of-the-art PU methods.

Download Full-text