LiteMuL: A Lightweight On-Device Sequence Tagger using Multi-task Learning

Named entity recognition (NER) often suffers from lack of annotation data. Multi-domain and multi-task learning solve this problem in some degree. However, previous multi-domain and multi-task learning are often studied in English. In the other part, multi-domain and multi-task learning are often researched independently. In this manuscript, we first summarize the previous works of multi-domain and multi-task learning in NER. Then, we introduce the multi-domain and multi-task learning in Chinese NER. Finally, we explore the universal models between multi-domain and multi-task learning. Experiments show that the universal models can be used in Chinese NER and outperform the baseline model.

Objective. Deep learning (DL)-based automated contouring and treatment planning has been proven to improve the efficiency and accuracy of radiotherapy. However, conventional radiotherapy treatment planning process has the automated contouring and treatment planning as separate tasks. Moreover in DL, the contouring and dose prediction tasks for automated treatment planning are done independently.Approach. In this study, we applied the multi-task learning (MTL) approach in order to seamlessly integrate automated contouring and voxel-based dose prediction tasks, as MTL can leverage common information between the two tasks and be able to increase the efficiency of the automated tasks. We developed our MTL framework using the two datasets: in-house prostate cancer dataset and the publicly available head and neck cancer dataset, OpenKBP.Main results. Compared to the sequential DL contouring and treatment planning tasks, our proposed method using MTL improved the mean absolute difference of dose volume histogram metrics of prostate and head and neck sites by 19.82% and 16.33%, respectively. Our MTL model for automated contouring and dose prediction tasks demonstrated enhanced dose prediction performance while maintaining or sometimes even improving the contouring accuracy. Compared to the baseline automated contouring model with the Dice score coefficients of 0.818 for prostate and 0.674 for head and neck datasets, our MTL approach achieved average scores of 0.824 and 0.716 for these datasets, respectively.Significance. Our study highlights the potential of the proposed automated contouring and planning using MTL to support the development of efficient and accurate automated treatment planning for radiotherapy.

Imitation learning is a promising approach to acquiring autonomous driving policies by mimicking human driver behaviors. However, a major drawback of existing driving policies derived from imitation learning is their proneness to capturing spurious correlations, owing to the lack of an explicit causal model. Deploying such policies in unpredictable real-world environments poses severe risks, as spurious correlations may result in flawed decisions that compromise safety. To tackle this challenge, we introduce a novel approach called Multi-Task Invariant Representation Imitation Learning (MIRIL). MIRIL combines invariant learning with imitation learning to identify cross-environment invariant causal representations from driving demonstrations in various scenarios. These representations are then fed into multiple downstream branches for multi-task learning, including policy learning, perception prediction, invariant representation learning, and transition dynamics learning. Through the multi-task learning approach, the model not only makes consistent driving decisions across different environments but also perceives the vehicle's surroundings, thereby improving adaptability and robustness in diverse driving conditions. This enables MIRIL to effectively handle a wide range of driving scenarios, ensuring safety and efficiency. Supported by clear metrics, this paper details our comprehensive experimental setup, including datasets, benchmarks, and comparative analyses, underscoring the capability of MIRIL to significantly boost system generalization and excel in decision-making significantly.

Neural-based multi-task learning has been successfully used in many real-world large-scale applications such as recommendation systems. For example, in movie recommendations, beyond providing users movies which they tend to purchase and watch, the system might also optimize for users liking the movies afterwards. With multi-task learning, we aim to build a single model that learns these multiple goals and tasks simultaneously. However, the prediction quality of commonly used multi-task models is often sensitive to the relationships between tasks. It is therefore important to study the modeling tradeoffs between task-specific objectives and inter-task relationships. In this work, we propose a novel multi-task learning approach, Multi-gate Mixture-of-Experts (MMoE), which explicitly learns to model task relationships from data. We adapt the Mixture-of-Experts (MoE) structure to multi-task learning by sharing the expert submodels across all tasks, while also having a gating network trained to optimize each task. To validate our approach on data with different levels of task relatedness, we first apply it to a synthetic dataset where we control the task relatedness. We show that the proposed approach performs better than baseline methods when the tasks are less related. We also show that the MMoE structure results in an additional trainability benefit, depending on different levels of randomness in the training data and model initialization. Furthermore, we demonstrate the performance improvements by MMoE on real tasks including a binary classification benchmark, and a large-scale content recommendation system at Google.

Joint predictions of multi-modal ride-hailing demands: A deep multi-task multi-graph learning-based approach

Software vulnerabilities seriously affect the security of computing systems and they are continuously disclosed and reported. When documenting software vulnerabilities, characterizing the severity, exploitability and impact of a vulnerability is critical for effective triaging and management of software vulnerabilities. Faced with ever-growing number of new vulnerabilities, we observe a significant lag between the disclosure of a vulnerability and the specification of its characteristics. This lag calls for automated, reliable assessment of vulnerability characteristics to assist security analysts in allocating their limited efforts to potentially most serious vulnerabilities. Existing automated techniques for vulnerability assessment require hand-crafted features and balanced data, and consider each specific characteristic independently at a time. In this paper, we propose a multi-task machine learning approach for the joint prediction of multiple vulnerability characteristics based on the vulnerability descriptions. Our approach gets rid of the requirement of balanced data, and it relies on neural networks that learn to extract features from training data. Using the large-scale vulnerability data in the Common Vulnerabilities and Exposure(CVE) database, we conduct extensive experiments to compare different configurations of neural network feature extractors, study the impact of multi-task learning versus independent-task learning, and investigate the performance of our approach for predicting the characteristics of newly disclosed vulnerabilities and the minimum requirement of historical vulnerability data for training reliable prediction model.

Depression is a common mood disorder that causes severe medical problems and interferes negatively with daily life. Identifying human behavior patterns that are predictive or indicative of depressive disorder is important. Clinical diagnosis of depression relies on costly clinician assessment using survey instruments which may not objectively reflect the fluctuation of daily behavior. Self-administered surveys, such as the Quick Inventory of Depressive Symptomatology (QIDS) commonly used to monitor depression, may show disparities from clinical decision. Smartphones provide easy access to many behavioral parameters, and Fitbit wrist bands are becoming another important tool to assess variables such as heart rates and sleep efficiency that are complementary to smartphone sensors. However, data used to identify depression indicators have been limited to a single platform either iPhone, or Android, or Fitbit alone due to the variation in their methods of data collection. The present work represents a large-scale effort to collect and integrate data from mobile phones, wearable devices, and self reports in depression analysis by designing a new machine learning approach. This approach constructs sparse mappings from sensing variables collected by various tools to two separate targets: self-reported QIDS scores and clinical assessment of depression severity. We propose a so-called heterogeneous multi-task feature learning method that jointly builds inference models for related tasks but of different types including classification and regression tasks. The proposed method was evaluated using data collected from 103 college students and could predict the QIDS score with an R2 reaching 0.44 and depression severity with an F1-score as high as 0.77. By imposing appropriate regularizers, our approach identified strong depression indicators such as time staying at home and total time asleep.

Plan recognition is a key component of player modeling. Player plan recognition focuses on modeling how and when players select goals and formulate action sequences to achieve their goals during gameplay. By occasionally asking players to describe their plans, it is possible to devise robust plan recognition models that jointly reason about player goals and action sequences in coordination with player input. In this work, we present a player plan recognition framework that leverages data from player interactions with a planning support tool embedded in an educational game for middle school science education, CRYSTAL ISLAND. Players are prompted to use the planning tool to describe their goals and planned actions in CRYSTAL ISLAND. We use this data to devise data-driven player plan recognition models using multi-label multi-task learning. Specifically, we compare single-task and multi-task learning approaches for both goal prediction and action sequence prediction. Results indicate that multi-task learning yields significant benefits for action sequence prediction. Additionally, we find that incorporating automated detectors of plan completion in plan recognition models improves predictive performance in both tasks.

Multitask learning has shown promising performance in learning multiple related tasks simultaneously, and variants of model architectures have been proposed, especially for supervised classification problems. One goal of multitask learning is to extract a good representation that sufficiently captures the relevant part of the input about the output for each learning task. To achieve this objective, in this paper we design a multitask learning architecture based on the observation that correlations exist between outputs of some related tasks (e.g. entity recognition and relation extraction tasks), and they reflect the relevant features that need to be extracted from the input. As outputs are unobserved, our proposed model exploits task predictions in lower layers of the neural model, also referred to as early predictions in this work. But we control the injection of early predictions to ensure that we extract good task-specific representations for classification. We refer to this model as a Progressive Multitask learning model with Explicit Interactions (PMEI). Extensive experiments on multiple benchmark datasets produce state-of-the-art results on the joint entity and relation extraction task.

Two inductive knowledge transfer approaches - multitask learning (MTL) and Feature Net (FN) - have been used to build predictive neural networks (ASNN) and PLS models for 11 types of tissue-air partition coefficients (TAPC). Unlike conventional single-task learning (STL) modeling focused only on a single target property without any relations to other properties, in the framework of inductive transfer approach, the individual models are viewed as nodes in the network of interrelated models built in parallel (MTL) or sequentially (FN). It has been demonstrated that MTL and FN techniques are extremely useful in structure-property modeling on small and structurally diverse data sets, when conventional STL modeling is unable to produce any predictive model. The predictive STL individual models were obtained for 4 out of 11 TAPC, whereas application of inductive knowledge transfer techniques resulted in models for 9 TAPC. Differences in prediction performances of the models as a function of the machine-learning method, and of the number of properties simultaneously involved in the learning, has been discussed.

Challenges and Opportunities of Using Transformer-Based Multi-Task Learning in NLP Through ML Lifecycle: A Position Paper

Semi-supervised named entity recognition in multi-level contexts

Convolutional neural networks (CNN) have been used successfully in solving many challenging visual perception tasks facing mobile robots and self-driving cars. To facilitate deploying such models on embedded hardware onboard mobile robots that have limited resources, multitask learning approaches have become common. In typically used multitask learning, a shared encoder network extracts features from inputs whereas multiple task-specific decoders transform these features into their target output. However, properly combining different tasks’ losses into the final network loss such that each task is making progress learning is a major challenge in these approaches. In this article, we present an innovative approach to extend a typical single-task network with the capability of performing two tasks without multiple decoders, i.e., a single-stream two-task network. The two output tasks are semantic segmentation and monocular depth prediction which are essential tasks in visual perception for autonomous driving. The method is centered on solving semantic segmentation with a regression loss function rather than a classification one. With our approach, we seize multitask learning benefits of reduced overhead and enhanced generalization while alleviating the need to balance different loss functions. Experimental evaluations with baseline single tasks and a multitask network are presented.

This study introduces a framework aimed at bolstering the effectiveness of deep learning models in image classification tasks, with a specialized focus on diabetic retinopathy detection. The proposed strategy employs transfer and multitask learning methods for the consistent extraction of insights and refinement of pre-established models. The capability of multitask learning to simultaneously acquire knowledge from numerous related tasks, leveraging shared information and inter-dependencies, leads to considerable performance boosts. Integrated into this process is the feature guideline learning task, which empowers the model to have an attention to abnormal areas. This task is initially trained on a small collection of labeled images to develop a custom pre-trained model, which later forms the basis for subsequent classification tasks. The effectiveness of this proposed approach is validated through tests on various data sets. It has been proven that our method contributes considerable advancements in image classification performance. This research offers valuable perspectives on the detection and classification of diabetic retinopathy, providing solutions to problems originating from limited data, and shows promising potential for improving patient care.

Data sparsity is a big challenge for collaborative filtering. This problem becomes more serious if the dataset is newly created and has even fewer ratings. By sharing knowledge among different datasets, multi-task learning is a promising technique to address this issue. Most prior work methods directly share objects (users or items) across different datasets. However, object identities and correspondences may not be known in many cases. We extend the previous work of Bayesian matrix factorization with Dirichlet process mixture into a multi-task learning approach by sharing latent parameters among different tasks. Our method does not require object identities and thus is more widely applicable. The proposed model is fully non-parametric in that the dimension of latent feature vectors is automatically determined. Inference is performed using the variational Bayesian algorithm, which is much faster than Gibbs sampling used by most other related Bayesian methods.