In most real-world large-scale online applications (e.g., e-commerce or finance), customer acquisition is usually a multi-step conversion process of audiences. For example, an impression->click->purchase process is usually performed of audiences for e-commerce platforms. However, it is more difficult to acquire customers in financial advertising (e.g., credit card advertising) than in traditional advertising. On the one hand, the audience multi-step conversion path is longer. On the other hand, the positive feedback is sparser (class imbalance) step by step, and it is difficult to obtain the final positive feedback due to the delayed feedback of activation. Multi-task learning is a typical solution in this direction. While considerable multi-task efforts have been made in this direction, a long-standing challenge is how to explicitly model the long-path sequential dependence among audience multi-step conversions for improving the end-to-end conversion. In this paper, we propose an Adaptive Information Transfer Multi-task (AITM) framework, which models the sequential dependence among audience multi-step conversions via the Adaptive Information Transfer (AIT) module. The AIT module can adaptively learn what and how much information to transfer for different conversion stages. Besides, by combining the Behavioral Expectation Calibrator in the loss function, the AITM framework can yield more accurate end-to-end conversion identification. The proposed framework is deployed in Meituan app, which utilizes it to real-timely show a banner to the audience with a high end-to-end conversion rate for Meituan Co-Branded Credit Cards. Offline experimental results on both industrial and public real-world datasets clearly demonstrate that the proposed framework achieves significantly better performance compared with state-of-the-art baselines.
In multistage manufacturing systems, modeling multiple quality indices based on the process sensing variables is important. However, the classic modeling technique predicts each quality variable one at a time, which fails to consider the correlation within or between stages. We propose a deep multistage multi-task learning framework to jointly predict all output sensing variables in a unified end-to-end learning framework according to the sequential system architecture in the MMS. Our numerical studies and real case study have shown that the new model has a superior performance compared to many benchmark methods as well as great interpretability through developed variable selection techniques.
The capability to automatically detect human stress can benefit artificial intelligent agents involved in affective computing and human-computer interaction. Stress and emotion are both human affective states, and stress has proven to have important implications on the regulation and expression of emotion. Although a series of methods have been established for multimodal stress detection, limited steps have been taken to explore the underlying inter-dependence between stress and emotion. In this work, we investigate the value of emotion recognition as an auxiliary task to improve stress detection. We propose MUSER -- a transformer-based model architecture and a novel multi-task learning algorithm with speed-based dynamic sampling strategy. Evaluations on the Multimodal Stressed Emotion (MuSE) dataset show that our model is effective for stress detection with both internal and external auxiliary tasks, and achieves state-of-the-art results.
We present an approach for encoding visual task relationships to improve model performance in an Unsupervised Domain Adaptation (UDA) setting. Semantic segmentation and monocular depth estimation are shown to be complementary tasks; in a multi-task learning setting, a proper encoding of their relationships can further improve performance on both tasks. Motivated by this observation, we propose a novel Cross-Task Relation Layer (CTRL), which encodes task dependencies between the semantic and depth predictions. To capture the cross-task relationships, we propose a neural network architecture that contains task-specific and cross-task refinement heads. Furthermore, we propose an Iterative Self-Learning (ISL) training scheme, which exploits semantic pseudo-labels to provide extra supervision on the target domain. We experimentally observe improvements in both tasks' performance because the complementary information present in these tasks is better captured. Specifically, we show that: (1) our approach improves performance on all tasks when they are complementary and mutually dependent; (2) the CTRL helps to improve both semantic segmentation and depth estimation tasks performance in the challenging UDA setting; (3) the proposed ISL training scheme further improves the semantic segmentation performance. The implementation is available at https://github.com/susaha/ctrl-uda.
In this paper, at first, the impact of ImageNet pre-training on fine-grained Facial Emotion Recognition (FER) is investigated which shows that when enough augmentations on images are applied, training from scratch provides better result than fine-tuning on ImageNet pre-training. Next, we propose a method to improve fine-grained and in-the-wild FER, called Hybrid Multi-Task Learning (HMTL). HMTL uses Self-Supervised Learning (SSL) as an auxiliary task during classical Supervised Learning (SL) in the form of Multi-Task Learning (MTL). Leveraging SSL during training can gain additional information from images for the primary fine-grained SL task. We investigate how proposed HMTL can be used in the FER domain by designing two customized version of common pre-text task techniques, puzzling and in-painting. We achieve state-of-the-art results on the AffectNet benchmark via two types of HMTL, without utilizing pre-training on additional data. Experimental results on the common SSL pre-training and proposed HMTL demonstrate the difference and superiority of our work. However, HMTL is not only limited to FER domain. Experiments on two types of fine-grained facial tasks, i.e., head pose estimation and gender recognition, reveals the potential of using HMTL to improve fine-grained facial representation.
We propose an ensemble model for predicting the lexical complexity of words and multiword expressions (MWEs). The model receives as input a sentence with a target word or MWEand outputs its complexity score. Given that a key challenge with this task is the limited size of annotated data, our model relies on pretrained contextual representations from different state-of-the-art transformer-based language models (i.e., BERT and RoBERTa), and on a variety of training methods for further enhancing model generalization and robustness:multi-step fine-tuning and multi-task learning, and adversarial training. Additionally, we propose to enrich contextual representations by adding hand-crafted features during training. Our model achieved competitive results and ranked among the top-10 systems in both sub-tasks.
Automated lobar segmentation allows regional evaluation of lung disease and is important for diagnosis and therapy planning. Advanced statistical workflows permitting such evaluation is a needed area within respiratory medicine; their adoption remains slow, with poor workflow accuracy. Diseased lung regions often produce high-density zones on CT images, limiting an algorithm's execution to specify damaged lobes due to oblique or lacking fissures. This impact motivated developing an improved machine learning method to segment lung lobes that utilises tracheobronchial tree information to enhance segmentation accuracy through the algorithm's spatial familiarity to define lobar extent more accurately. The method undertakes parallel segmentation of lobes and auxiliary tissues simultaneously by employing multi-task learning (MTL) in conjunction with V-Net-attention, a popular convolutional neural network in the imaging realm. In keeping with the model's adeptness for better generalisation, high performance was retained in an external dataset of patients with four distinct diseases: severe lung cancer, COVID-19 pneumonitis, collapsed lungs and Chronic Obstructive Pulmonary Disease (COPD), even though the training data included none of these cases. The benefit of our external validation test is specifically relevant since our choice includes those patients who have diagnosed lung disease with associated radiological abnormalities. To ensure equal rank is given to all segmentations in the main task we report the following performance (Dice score) on a per-segment basis: normal lungs 0.97, COPD 0.94, lung cancer 0.94, COVID-19 pneumonitis 0.94 and collapsed lung 0.92, all at p<0.05. Even segmenting lobes with large deformations on CT images, the model maintained high accuracy. The approach can be readily adopted in the clinical setting as a robust tool for radiologists.
This paper considers the problem of differentially private semi-supervised transfer and multi-task learning. The notion of \emph{membership-mapping} has been developed using measure theory basis to learn data representation via a fuzzy membership function. An alternative conception of deep autoencoder, referred to as \emph{Conditionally Deep Membership-Mapping Autoencoder (CDMMA)}, is considered for transferrable deep learning. Under practice-oriented settings, an analytical solution for the learning of CDMMA can be derived by means of variational optimization. The paper proposes a transfer and multi-task learning approach that combines CDMMA with a tailored noise adding mechanism to achieve a given level of privacy-loss bound with the minimum perturbation of the data. Numerous experiments were carried out using MNIST, USPS, Office, and Caltech256 datasets to verify the competitive robust performance of the proposed methodology.
Multi-Task Learning has emerged as a methodology in which multiple tasks are jointly learned by a shared learning algorithm, such as a DNN. MTL is based on the assumption that the tasks under consideration are related; therefore it exploits shared knowledge for improving performance on each individual task. Tasks are generally considered to be homogeneous, i.e., to refer to the same type of problem. Moreover, MTL is usually based on ground truth annotations with full, or partial overlap across tasks. In this work, we deal with heterogeneous MTL, simultaneously addressing detection, classification & regression problems. We explore task-relatedness as a means for co-training, in a weakly-supervised way, tasks that contain little, or even non-overlapping annotations. Task-relatedness is introduced in MTL, either explicitly through prior expert knowledge, or through data-driven studies. We propose a novel distribution matching approach, in which knowledge exchange is enabled between tasks, via matching of their predictions' distributions. Based on this approach, we build FaceBehaviorNet, the first framework for large-scale face analysis, by jointly learning all facial behavior tasks. We develop case studies for: i) continuous affect estimation, action unit detection, basic emotion recognition; ii) attribute detection, face identification. We illustrate that co-training via task relatedness alleviates negative transfer. Since FaceBehaviorNet learns features that encapsulate all aspects of facial behavior, we conduct zero-/few-shot learning to perform tasks beyond the ones that it has been trained for, such as compound emotion recognition. By conducting a very large experimental study, utilizing 10 databases, we illustrate that our approach outperforms, by large margins, the state-of-the-art in all tasks and in all databases, even in these which have not been used in its training.
People ``understand'' the world via vision, hearing, tactile, and also the past experience. Human experience can be learned through normal learning (we call it explicit knowledge), or subconsciously (we call it implicit knowledge). These experiences learned through normal learning or subconsciously will be encoded and stored in the brain. Using these abundant experience as a huge database, human beings can effectively process data, even they were unseen beforehand. In this paper, we propose a unified network to encode implicit knowledge and explicit knowledge together, just like the human brain can learn knowledge from normal learning as well as subconsciousness learning. The unified network can generate a unified representation to simultaneously serve various tasks. We can perform kernel space alignment, prediction refinement, and multi-task learning in a convolutional neural network. The results demonstrate that when implicit knowledge is introduced into the neural network, it benefits the performance of all tasks. We further analyze the implicit representation learnt from the proposed unified network, and it shows great capability on catching the physical meaning of different tasks. The source code of this work is at : https://github.com/WongKinYiu/yolor.
Named entity recognition (NER) has been studied extensively and the earlier algorithms were based on sequence labeling like Hidden Markov Models (HMM) and conditional random fields (CRF). These were followed by neural network based deep learning models. Recently, BERT has shown new state of the art accuracy in sequence labeling tasks like NER. In this short article, we study various approaches to task specific NER. Task specific NER has two components - identifying the intent of a piece of text (like search queries), and then labeling the query with task specific named entities. For example, we consider the task of labeling Target store locations in a search query (which could be entered in a search box or spoken in a device like Alexa or Google Home). Store locations are highly ambiguous and sometimes it is difficult to differentiate between say a location and a non-location. For example, "pickup my order at orange store" has "orange" as the store location, while "buy orange at target" has "orange" as a fruit. We explore this difficulty by doing multi-task learning which we call global to local transfer of information. We jointly learn the query intent (i.e. store lookup) and the named entities by using multiple loss functions in our BERT based model and find interesting results.
There is a wide variety of speech processing tasks ranging from extracting content information from speech signals to generating speech signals. For different tasks, model networks are usually designed and tuned separately. If a universal model can perform multiple speech processing tasks, some tasks might be improved with the related abilities learned from other tasks. The multi-task learning of a wide variety of speech processing tasks with a universal model has not been studied. This paper proposes a universal modularized model, SpeechNet, which treats all speech processing tasks into a speech/text input and speech/text output format. We select five essential speech processing tasks for multi-task learning experiments with SpeechNet. We show that SpeechNet learns all of the above tasks, and we further analyze which tasks can be improved by other tasks. SpeechNet is modularized and flexible for incorporating more modules, tasks, or training approaches in the future. We release the code and experimental settings to facilitate the research of modularized universal models and multi-task learning of speech processing tasks.
Medical image registration and segmentation are two of the most frequent tasks in medical image analysis. As these tasks are complementary and correlated, it would be beneficial to apply them simultaneously in a joint manner. In this paper, we formulate registration and segmentation as a joint problem via a Multi-Task Learning (MTL) setting, allowing these tasks to leverage their strengths and mitigate their weaknesses through the sharing of beneficial information. We propose to merge these tasks not only on the loss level, but on the architectural level as well. We studied this approach in the context of adaptive image-guided radiotherapy for prostate cancer, where planning and follow-up CT images as well as their corresponding contours are available for training. The study involves two datasets from different manufacturers and institutes. The first dataset was divided into training (12 patients) and validation (6 patients), and was used to optimize and validate the methodology, while the second dataset (14 patients) was used as an independent test set. We carried out an extensive quantitative comparison between the quality of the automatically generated contours from different network architectures as well as loss weighting methods. Moreover, we evaluated the quality of the generated deformation vector field (DVF). We show that MTL algorithms outperform their Single-Task Learning (STL) counterparts and achieve better generalization on the independent test set. The best algorithm achieved a mean surface distance of $1.06 \pm 0.3$ mm, $1.27 \pm 0.4$ mm, $0.91 \pm 0.4$ mm, and $1.76 \pm 0.8$ mm on the validation set for the prostate, seminal vesicles, bladder, and rectum, respectively. The high accuracy of the proposed method combined with the fast inference speed, makes it a promising method for automatic re-contouring of follow-up scans for adaptive radiotherapy.
Learning of matrix-valued data has recently surged in a range of scientific and business applications. Trace regression is a widely used method to model effects of matrix predictors and has shown great success in matrix learning. However, nearly all existing trace regression solutions rely on two assumptions: (i) a known functional form of the conditional mean, and (ii) a global low-rank structure in the entire range of the regression function, both of which may be violated in practice. In this article, we relax these assumptions by developing a general framework for nonparametric trace regression models via structured sign series representations of high dimensional functions. The new model embraces both linear and nonlinear trace effects, and enjoys rank invariance to order-preserving transformations of the response. In the context of matrix completion, our framework leads to a substantially richer model based on what we coin as the "sign rank" of a matrix. We show that the sign series can be statistically characterized by weighted classification tasks. Based on this connection, we propose a learning reduction approach to learn the regression model via a series of classifiers, and develop a parallelable computation algorithm to implement sign series aggregations. We establish the excess risk bounds, estimation error rates, and sample complexities. Our proposal provides a broad nonparametric paradigm to many important matrix learning problems, including matrix regression, matrix completion, multi-task learning, and compressed sensing. We demonstrate the advantages of our method through simulations and two applications, one on brain connectivity study and the other on high-rank image completion.
User queries for a real-world dialog system may sometimes fall outside the scope of the system's capabilities, but appropriate system responses will enable smooth processing throughout the human-computer interaction. This paper is concerned with the user's intent, and focuses on out-of-scope intent classification in dialog systems. Although user intents are highly correlated with the application domain, few studies have exploited such correlations for intent classification. Rather than developing a two-stage approach that first classifies the domain and then the intent, we propose a hierarchical multi-task learning approach based on a joint model to classify domain and intent simultaneously. Novelties in the proposed approach include: (1) sharing supervised out-of-scope signals in joint modeling of domain and intent classification to replace a two-stage pipeline; and (2) introducing a hierarchical model that learns the intent and domain representations in the higher and lower layers respectively. Experiments show that the model outperforms existing methods in terms of accuracy, out-of-scope recall and F1. Additionally, threshold-based post-processing further improves performance by balancing precision and recall in intent classification.
Building ASR models across many languages is a challenging multi-task learning problem due to large variations and heavily unbalanced data. Existing work has shown positive transfer from high resource to low resource languages. However, degradations on high resource languages are commonly observed due to interference from the heterogeneous multilingual data and reduction in per-language capacity. We conduct a capacity study on a 15-language task, with the amount of data per language varying from 7.6K to 53.5K hours. We adopt GShard [1] to efficiently scale up to 10B parameters. Empirically, we find that (1) scaling the number of model parameters is an effective way to solve the capacity bottleneck - our 500M-param model already outperforms monolingual baselines and scaling it to 1B and 10B brought further quality gains; (2) larger models are not only more data efficient, but also more efficient in terms of training cost as measured in TPU days - the 1B-param model reaches the same accuracy at 34% of training time as the 500M-param model; (3) given a fixed capacity budget, adding depth works better than width and large encoders do better than large decoders; (4) with continuous training, they can be adapted to new languages and domains.
Recently, researchers have utilized neural networks to accurately solve partial differential equations (PDEs), enabling the mesh-free method for scientific computation. Unfortunately, the network performance drops when encountering a high nonlinearity domain. To improve the generalizability, we introduce the novel approach of employing multi-task learning techniques, the uncertainty-weighting loss and the gradients surgery, in the context of learning PDE solutions. The multi-task scheme exploits the benefits of learning shared representations, controlled by cross-stitch modules, between multiple related PDEs, which are obtainable by varying the PDE parameterization coefficients, to generalize better on the original PDE. Encouraging the network pay closer attention to the high nonlinearity domain regions that are more challenging to learn, we also propose adversarial training for generating supplementary high-loss samples, similarly distributed to the original training distribution. In the experiments, our proposed methods are found to be effective and reduce the error on the unseen data points as compared to the previous approaches in various PDE examples, including high-dimensional stochastic PDEs.
Scene graph generation has emerged as an important problem in computer vision. While scene graphs provide a grounded representation of objects, their locations and relations in an image, they do so only at the granularity of proposal bounding boxes. In this work, we propose the first, to our knowledge, framework for pixel-level segmentation-grounded scene graph generation. Our framework is agnostic to the underlying scene graph generation method and address the lack of segmentation annotations in target scene graph datasets (e.g., Visual Genome) through transfer and multi-task learning from, and with, an auxiliary dataset (e.g., MS COCO). Specifically, each target object being detected is endowed with a segmentation mask, which is expressed as a lingual-similarity weighted linear combination over categories that have annotations present in an auxiliary dataset. These inferred masks, along with a novel Gaussian attention mechanism which grounds the relations at a pixel-level within the image, allow for improved relation prediction. The entire framework is end-to-end trainable and is learned in a multi-task manner with both target and auxiliary datasets.
Confidence scores are very useful for downstream applications of automatic speech recognition (ASR) systems. Recent works have proposed using neural networks to learn word or utterance confidence scores for end-to-end ASR. In those studies, word confidence by itself does not model deletions, and utterance confidence does not take advantage of word-level training signals. This paper proposes to jointly learn word confidence, word deletion, and utterance confidence. Empirical results show that multi-task learning with all three objectives improves confidence metrics (NCE, AUC, RMSE) without the need for increasing the model size of the confidence estimation module. Using the utterance-level confidence for rescoring also decreases the word error rates on Google's Voice Search and Long-tail Maps datasets by 3-5% relative, without needing a dedicated neural rescorer.
The lottery ticket hypothesis (LTH) claims that a deep neural network (i.e., ground network) contains a number of subnetworks (i.e., winning tickets), each of which exhibiting identically accurate inference capability as that of the ground network. Federated learning (FL) has recently been applied in LotteryFL to discover such winning tickets in a distributed way, showing higher accuracy multi-task learning than Vanilla FL. Nonetheless, LotteryFL relies on unicast transmission on the downlink, and ignores mitigating stragglers, questioning scalability. Motivated by this, in this article we propose a personalized and communication-efficient federated lottery ticket learning algorithm, coined CELL, which exploits downlink broadcast for communication efficiency. Furthermore, it utilizes a novel user grouping method, thereby alternating between FL and lottery learning to mitigate stragglers. Numerical simulations validate that CELL achieves up to 3.6% higher personalized task classification accuracy with 4.3x smaller total communication cost until convergence under the CIFAR-10 dataset.