A review on multi-task metric learning

Multi-task learning (MTL) is a machine learning paradigm that improves the performance of each task by exploiting useful information contained in multiple related tasks. However, the relatedness of tasks can be exploited by attackers to launch data poisoning attacks, which has been demonstrated a big threat to single-task learning. In this paper, we provide the first study on the vulnerability of MTL. Specifically, we focus on multi-task relationship learning (MTRL) models, a popular subclass of MTL models where task relationships are quantized and are learned directly from training data. We formulate the problem of computing optimal poisoning attacks on MTRL as a bilevel program that is adaptive to arbitrary choice of target tasks and attacking tasks. We propose an efficient algorithm called PATOM for computing optimal attack strategies. PATOM leverages the optimality conditions of the subproblem of MTRL to compute the implicit gradients of the upper level objective function. Experimental results on real-world datasets show that MTRL models are very sensitive to poisoning attacks and the attacker can significantly degrade the performance of target tasks, by either directly poisoning the target tasks or indirectly poisoning the related tasks exploiting the task relatedness. We also found that the tasks being attacked are always strongly correlated, which provides a clue for defending against such attacks.

Multi-task feature learning has been proposed to improve the generalization performance by learning the shared features among multiple related tasks and it has been successfully applied to many real-world problems in machine learning, data mining, computer vision and bioinformatics. Most existing multi-task feature learning models simply assume a common noise level for all tasks, which may not be the case in real applications. Recently, a Calibrated Multivariate Regression (CMR) model has been proposed, which calibrates different tasks with respect to their noise levels and achieves superior prediction performance over the non-calibrated one. A major challenge is how to solve the CMR model efficiently as it is formulated as a composite optimization problem consisting of two non-smooth terms. In this paper, we propose a variant of the calibrated multi-task feature learning formulation by including a squared norm regularizer. We show that the dual problem of the proposed formulation is a smooth optimization problem with a piecewise sphere constraint. The simplicity of the dual problem enables us to develop fast dual optimization algorithms with low per-iteration cost. We also provide a detailed convergence analysis for the proposed dual optimization algorithm. Empirical studies demonstrate that, the dual optimization algorithm quickly converges and it is much more efficient than the primal optimization algorithm. Moreover, the calibrated multi-task feature learning algorithms with and without the squared norm regularizer achieve similar prediction performance and both outperform the non-calibrated ones. Thus, the proposed variant not only enables us to develop fast optimization algorithms, but also keeps the superior prediction performance of the calibrated multi-task feature learning over the non-calibrated one.

Multi-task learning has been established as an important machine learning framework for leveraging shared knowledge among multiple different but related tasks, with the generalization performance of models enhanced. As a promising learning paradigm, multi-task learning has been widely adopted by various real-world applications, such as recommendation systems. Multi-gate Mixture-of-Experts (MMoE), a well-received multi-task learning method in industry, based on the classic and inspiring Mixture-of-Experts (MoE) structure, explicitly models task relationships and learns task-specific functionalities, generating significant improvements. However, in our applications, negative transfer, which confuses considerable existing multi-task learning methods, is still observed to happen to MMoE. In this paper, an in-depth empirical investigation into negative transfer is launched. And it reveals that, incompetent experts, which play fundamental roles under the learning framework of MoE, are the key technique bottleneck. To tackle this dilemma, we propose the Calibrated Mixture of Insightful Experts (CMoIE), with three novel modules (Conflict Resolution, Expert Communication, and Mixture Calibration), customed for multi-task learning. Hence a group of insightful experts are constructed with enhanced diversity, communication and specialization. To validate the proposed method CMoIE, experiments are conducted on three public datasets and one real-world click-through-rate prediction dataset we construct based on traffic logs collected from a large-scale online product recommendation system. Our approach yields best performance across all of these benchmarks, demonstrating the superiority of it.

In machine learning, the relatedness across multiple tasks is usually complex and entangled. Due to dataset bias, the relatedness among tasks might be distorted and mislead the training of the models with solid learning ability, such as the multi-task neural networks. In this paper, we propose the idea of Relatedness Refinement Multi-Task Learning (RRMTDL) by introducing adversarial learning in the multi-task deep neural network to tackle the problem. The RRMTDL deep learning model restrains the misleading relatedness task by adversarial training and extracts information sharing across tasks with valuable relatedness. With RRMTDL, multi-task deep learning can enhance the task-specific representation for the major tasks by excluding the misleading relatedness. We design tests with various combinations of task-relatedness to validate the proposed model. Experimental results show that the RRMTDL model can effectively refine the task relatedness and prominently outperform other multi-task deep learning models in datasets with entangled task labels.

Multi-task learning (MTL) supports simultaneous training over multiple related tasks and learns the shared representation. While improving the generalization ability of training on a single task, MTL has higher privacy risk than traditional single-task learning because more sensitive information is extracted and learned in a correlated manner. Unfortunately, very few works have attempted to address the privacy risks posed by MTL. In this paper, we first investigate such risk by designing model extraction attack (MEA) and membership inference attack (MIA) in MTL. Then we evaluate the privacy risks on six MTL model architectures and two popular MTL datasets, whose results show that both the number of tasks and the complexity of training data play an important role in the attack performance. Our investigation shows that MTL is more vulnerable than traditional single-task learning under both attacks.

In recent years, multitask learning has turned out to be of great success in various applications. Though single task model training has promised great results throughout these years, it ignores valuable information which might help in estimating better learning parameters. By virtue of learning-related tasks, multitask learning has been able to generalize the models even better. In this paper, we try to enhance the feature mapping of multitask deep learning models by sharing features among related tasks and inductive transfer learning. We have explored the learning of task relationship among various tasks for acquiring more benefits from multitask learning. The proposed enhanced model is compared with two state-of-the-art multitask deep learning models, namely Hyperface and FaceNet. The results show better performance of the proposed model in predicting age, gender and ethnicity on the UTK face dataset.KeywordsNeural networksMultitask deep learningDeep learningRecurrent neural networksConvolution neural networks

Patients in the intensive care unit (ICU) require constant and close supervision. To assist clinical staff in this task, hospitals use monitoring systems that trigger audiovisual alarms if their algorithms indicate that a patient's condition may be worsening. However, current monitoring systems are extremely sensitive to movement artefacts and technical errors. As a result, they typically trigger hundreds to thousands of false alarms per patient per day - drowning the important alarms in noise and adding to the exhaustion of clinical staff. In this setting, data is abundantly available, but obtaining trustworthy annotations by experts is laborious and expensive. We frame the problem of false alarm reduction from multivariate time series as a machine-learning task and address it with a novel multitask network architecture that utilises distant supervision through multiple related auxiliary tasks in order to reduce the number of expensive labels required for training. We show that our approach leads to significant improvements over several state-of-the-art baselines on real-world ICU data and provide new insights on the importance of task selection and architectural choices in distantly supervised multitask learning.

Multi-task learning (MTL) exhibits improved performance in many problems in reality by utilizing the intrinsic features among multiple related tasks. In this paper, the problem of water quality monitoring is considered as multi-task learning, in which different tasks correspond to changes caused by new environment or different spectrometers. An improved learning model is presented describing the relationship between wavelengths and pollutant concentration as well as capturing the task relationships with a low-rank shared structure. Under the assumption that different tasks share some common wavelengths, an optimization problem is proposed with the predictors affected by these features and their corresponding coefficients that vary in different tasks. An alternating minimization algorithm is proposed to solve this problem. Experimental results demonstrate the effectiveness of the proposed algorithm in application.

Laser-based metal additive manufacturing (MAM), particularly laser powder bed fusion (LPBF) and laser directed energy deposition (L-DED), is a transformative technology enabling the production of complex, high-performance metal components with precision and minimal waste. However, its wider adoption is hindered by challenges in ensuring repeatability and consistency due to the complex, dynamic nature of the process. Closed-loop systems with in-situ monitoring and control have gained attention for quality assurance, where machine learning (ML) models are crucial for correlating sensor data with process state diagnosis and facilitating control actions. This review adopts an ML perspective to elucidate the role of ML in closed-loop control systems. It first outlines sensor data types and preprocessing methods used as ML datasets, then discusses roles and applications of ML models for in-situ monitoring, highlighting limitations and potential solutions. An ML-based closed-loop control framework is also presented, with emphasis on the integration of ML in the control component. Finally, future development directions of advanced ML approaches are envisioned. This review provides new insights into enhancing the role of ML models in laser-based MAM closed-loop control systems, serving as a crucial aid in the development of intelligent control systems.

Multi-task and multi-view learning of user state

Recent research in disaster informatics demonstrates a practical and important use case of artificial intelligence to save human lives and suffering during natural disasters based on social media contents (text and images). While notable progress has been made using texts, research on exploiting the images remains relatively under-explored. To advance image-based approaches, we propose MEDIC (https://crisisnlp.qcri.org/medic/index.html), which is the largest social media image classification dataset for humanitarian response consisting of 71,198 images to address four different tasks in a multi-task learning setup. This is the first dataset of its kind: social media images, disaster response, and multi-task learning research. An important property of this dataset is its high potential to facilitate research on multi-task learning, which recently receives much interest from the machine learning community and has shown remarkable results in terms of memory, inference speed, performance, and generalization capability. Therefore, the proposed dataset is an important resource for advancing image-based disaster management and multi-task machine learning research. We experiment with different deep learning architectures and report promising results, which are above the majority baselines for all tasks. Along with the dataset, we also release all relevant scripts (https://github.com/firojalam/medic).

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.

Navigating through the Maze of Homogeneous Catalyst Design with Machine Learning

Multi-Task Learning (MTL) for text classification takes advantage of the data to train a single shared model with multiple task-specific layers on multiple related classification tasks to improve its generalization performance. We choose pre-trained language models (BERT-family) as the shared part of this architecture. Although they have achieved noticeable performance in different downstream NLP tasks, their performance in an MTL setting for the biomedical domain is not thoroughly investigated. In this work, we investigate the performance of BERT-family models in different MTL settings with Open-I (radiology reports) and OHSUMED (PubMed abstracts) datasets. We introduce the MTLV (Multi-Task Learning Visualizer) library for building Multi-task learning-related architectures which use existing infrastructure (e.g., Hugging Face Transformers and MLflow Tracking). Following previous work in computer vision, we clustered tasks and trained a separate model on each cluster (Grouped Multi-Task Learning (GMTL)). Contextual representation of the class labels (Tasks) and their descriptions was used by the library as features to cluster the tasks. We observed that grouping tasks for training with few models (GMTL) outperforms the MTL also GMTL is computationally more efficient than the STL setting (a separate model is trained for each task).

<p><b>Dynamic flexible job shop scheduling (DFJSS) has received widespread attention from academia and industry due to its reflection in real-world scheduling applications such as order picking in the warehouse and the manufacturing industry. It requires complex routing and sequencing decisions under unpredicted dynamic events. Genetic programming, as ahyper-heuristic approach (GPHH), has been successfully applied to evolve scheduling heuristics for DFJSS automatically due to its flexible representation. Although GPHH has achieved certain success in solving the DFJSS problems, there are still some limitations for applying GPHH to DFJSS, particularly in terms of its training efficiency, large search space, search mechanism, and multitask solving ability. The overall goal of this thesis is to develop effective GPHH algorithmsto evolve scheduling heuristics for DFJSS efficiently. Different machine learning techniques, i.e., surrogate, feature selection, specialised genetic operator, and multitask learning, are incorporated in this thesis to tackle the limitations.</b></p> <p>First, this thesis develops a novel multi-fidelity based surrogate-assisted GPHH for DFJSS to improve the training efficiency of GPHH. Specifically, multi-fidelity based surrogate models are first designed by simplifying the problem to be solved. Then, an effective collaboration mechanism with knowledge transfer is proposed for utilising the advantages of the multifidelity based surrogate models to solve the problem. The results show that the proposed algorithm can dramatically reduce the computational cost of GPHH without sacrificing the performance in all the test scenarios. With the same training time, the proposed algorithm can achieve significantly better performance than its counterparts in most scenarios while no worse in others.</p> <p>Second, this thesis designs a novel two-stage GPHH framework with feature selection to evolve scheduling heuristics for DFJSS automatically. Based on this framework, this thesis further proposes to evolve scheduling heuristics with only the selected features by eliminating the unselected features properly. Specifically, individual adaptation strategies are proposed to generate individuals with only the selected features by utilising the information of both the selected features and the investigated individuals during the feature selection process. The results show that the proposed algorithm can successfully achieve scheduling heuristics with fewer unique features and smaller sizes, which tends to be more interpretable. In addition, the proposed algorithm can evolve comparable scheduling heuristic with that obtained by the traditional GPHH within a much shorter training time.</p> <p>Third, this thesis proposes a novel recombinative mechanism to provide guidance for GPHH based on the importance of subtrees to realise effective and adaptive recombination for parents to produce offspring. Two measures are proposed to measure the importance of all the subtrees of an individual. The first one is based on the frequency of features, and the second is based on the correlation between the behaviour of subtrees and the whole tree (i.e., an individual). The importance information is utilised to decide the crossover points for the parents. The proposed recombinative guidance mechanism attempts to improve the quality of offspring by preserving the promising building-blocks of one parent and incorporating good building-blocks from the other. The results show that the proposed algorithm based on the correlation importance measure performs better than the proposed algorithm based on the feature frequency importance measure. In addition, the proposed algorithm based on the correlation importance measure between the behaviour of subtrees significantly also outperforms the state-of-the-art algorithms on most tested scenarios.</p> <p>Last, this thesis proposes a multitask GPHH approach and a surrogate-assisted multitask GPHH approach to solving multiple DFJSS tasks simultaneously. First, an effective hyper-heuristic multitask algorithm is proposed by adapting the traditional evolutionary multitask algorithms based on the characteristics of GPHH. Second, this thesis develops a novel surrogate-assisted multitask GPHH approach to solving multiple DFJSS tasks by sharing useful knowledge between different DFJSS scheduling tasks. Specifically, the surrogate-assisted multitask GPHH algorithm employs the phenotypic characterisation technique to measure the behaviours of scheduling rules to build a surrogate for each task accordingly. The built surrogates are not only used to improve the efficiency of solving each single DFJSS task but also utilised for knowledge sharing between multiple DFJSS tasks in multitask learning. The results show that the proposed algorithm can significantly improve the quality of scheduling heuristics for all the test scenarios. The results also observe that the proposed algorithms manage to solve multiple tasks collaboratively in terms of the evolved scheduling heuristics for different tasks in a multitask scenario.</p>