The integration of additional side information to improve music source separation has been investigated numerous times, e.g., by adding features to the input or by adding learning targets in a multi-task learning scenario. These approaches, however, require additional annotations such as musical scores, instrument labels, etc. in training and possibly during inference. The available datasets for source separation do not usually provide these additional annotations. In this work, we explore transfer learning strategies to incorporate VGGish features with a state-of-the-art source separation model; VGGish features are known to be a very condensed representation of audio content and have been successfully used in many MIR tasks. We introduce three approaches to incorporate the features, including two latent space regularization methods and one naive concatenation method. Experimental results show that our proposed approaches improve several evaluation metrics for music source separation.
Speaker recognition is increasingly used in several everyday applications including smart speakers, customer care centers and other speech-driven analytics. It is crucial to accurately evaluate and mitigate biases present in machine learning (ML) based speech technologies, such as speaker recognition, to ensure their inclusive adoption. ML fairness studies with respect to various demographic factors in modern speaker recognition systems are lagging compared to other human-centered applications such as face recognition. Existing studies on fairness in speaker recognition systems are largely limited to evaluating biases at specific operating points of the systems, which can lead to false expectations of fairness. Moreover, there are only a handful of bias mitigation strategies developed for speaker recognition systems. In this paper, we systematically evaluate the biases present in speaker recognition systems with respect to gender across a range of system operating points. We also propose adversarial and multi-task learning techniques to improve the fairness of these systems. We show through quantitative and qualitative evaluations that the proposed methods improve the fairness of ASV systems over baseline methods trained using data balancing techniques. We also present a fairness-utility trade-off analysis to jointly examine fairness and the overall system performance. We show that although systems trained using adversarial techniques improve fairness, they are prone to reduced utility. On the other hand, multi-task methods can improve the fairness while retaining the utility. These findings can inform the choice of bias mitigation strategies in the field of speaker recognition.
The ability to predict the future trajectory of a patient is a key step toward the development of therapeutics for complex diseases such as Alzheimer's disease (AD). However, most machine learning approaches developed for prediction of disease progression are either single-task or single-modality models, which can not be directly adopted to our setting involving multi-task learning with high dimensional images. Moreover, most of those approaches are trained on a single dataset (i.e. cohort), which can not be generalized to other cohorts. We propose a novel multimodal multi-task deep learning model to predict AD progression by analyzing longitudinal clinical and neuroimaging data from multiple cohorts. Our proposed model integrates high dimensional MRI features from a 3D convolutional neural network with other data modalities, including clinical and demographic information, to predict the future trajectory of patients. Our model employs an adversarial loss to alleviate the study-specific imaging bias, in particular the inter-study domain shifts. In addition, a Sharpness-Aware Minimization (SAM) optimization technique is applied to further improve model generalization. The proposed model is trained and tested on various datasets in order to evaluate and validate the results. Our results showed that 1) our model yields significant improvement over the baseline models, and 2) models using extracted neuroimaging features from 3D convolutional neural network outperform the same models when applied to MRI-derived volumetric features.
While pretrained language models (PLMs) have been shown to possess a plethora of linguistic knowledge, the existing body of research has largely neglected extralinguistic knowledge, which is generally difficult to obtain by pretraining on text alone. Here, we contribute to closing this gap by examining geolinguistic knowledge, i.e., knowledge about geographic variation in language. We introduce geoadaptation, an intermediate training step that couples language modeling with geolocation prediction in a multi-task learning setup. We geoadapt four PLMs, covering language groups from three geographic areas, and evaluate them on five different tasks: fine-tuned (i.e., supervised) geolocation prediction, zero-shot (i.e., unsupervised) geolocation prediction, fine-tuned language identification, zero-shot language identification, and zero-shot prediction of dialect features. Geoadaptation is very successful at injecting geolinguistic knowledge into the PLMs: the geoadapted PLMs consistently outperform PLMs adapted using only language modeling (by especially wide margins on zero-shot prediction tasks), and we obtain new state-of-the-art results on two benchmarks for geolocation prediction and language identification. Furthermore, we show that the effectiveness of geoadaptation stems from its ability to geographically retrofit the representation space of the PLMs.
Large multilingual pretrained language models such as mBERT and XLM-RoBERTa have been found to be surprisingly effective for cross-lingual transfer of syntactic parsing models (Wu and Dredze 2019), but only between related languages. However, source and training languages are rarely related, when parsing truly low-resource languages. To close this gap, we adopt a method from multi-task learning, which relies on automated curriculum learning, to dynamically optimize for parsing performance on outlier languages. We show that this approach is significantly better than uniform and size-proportional sampling in the zero-shot setting.
While both extractive and generative readers have been successfully applied to the Question Answering (QA) task, little attention has been paid toward the systematic comparison of them. Characterizing the strengths and weaknesses of the two readers is crucial not only for making a more informed reader selection in practice but also for developing a deeper understanding to foster further research on improving readers in a principled manner. Motivated by this goal, we make the first attempt to systematically study the comparison of extractive and generative readers for question answering. To be aligned with the state-of-the-art, we explore nine transformer-based large pre-trained language models (PrLMs) as backbone architectures. Furthermore, we organize our findings under two main categories: (1) keeping the architecture invariant, and (2) varying the underlying PrLMs. Among several interesting findings, it is important to highlight that (1) the generative readers perform better in long context QA, (2) the extractive readers perform better in short context while also showing better out-of-domain generalization, and (3) the encoder of encoder-decoder PrLMs (e.g., T5) turns out to be a strong extractive reader and outperforms the standard choice of encoder-only PrLMs (e.g., RoBERTa). We also study the effect of multi-task learning on the two types of readers varying the underlying PrLMs and perform qualitative and quantitative diagnosis to provide further insights into future directions in modeling better readers.
The aspect-based sentiment analysis (ABSA) is a fine-grained task that aims to determine the sentiment polarity towards targeted aspect terms occurring in the sentence. The development of the ABSA task is very much hindered by the lack of annotated data. To tackle this, the prior works have studied the possibility of utilizing the sentiment analysis (SA) datasets to assist in training the ABSA model, primarily via pretraining or multi-task learning. In this article, we follow this line, and for the first time, we manage to apply the Pseudo-Label (PL) method to merge the two homogeneous tasks. While it seems straightforward to use generated pseudo labels to handle this case of label granularity unification for two highly related tasks, we identify its major challenge in this paper and propose a novel framework, dubbed as Dual-granularity Pseudo Labeling (DPL). Further, similar to PL, we regard the DPL as a general framework capable of combining other prior methods in the literature. Through extensive experiments, DPL has achieved state-of-the-art performance on standard benchmarks surpassing the prior work significantly.
Effectiveness and interpretability are two essential properties for trustworthy AI systems. Most recent studies in visual reasoning are dedicated to improving the accuracy of predicted answers, and less attention is paid to explaining the rationales behind the decisions. As a result, they commonly take advantage of spurious biases instead of actually reasoning on the visual-textual data, and have yet developed the capability to explain their decision making by considering key information from both modalities. This paper aims to close the gap from three distinct perspectives: first, we define a new type of multi-modal explanations that explain the decisions by progressively traversing the reasoning process and grounding keywords in the images. We develop a functional program to sequentially execute different reasoning steps and construct a new dataset with 1,040,830 multi-modal explanations. Second, we identify the critical need to tightly couple important components across the visual and textual modalities for explaining the decisions, and propose a novel explanation generation method that explicitly models the pairwise correspondence between words and regions of interest. It improves the visual grounding capability by a considerable margin, resulting in enhanced interpretability and reasoning performance. Finally, with our new data and method, we perform extensive analyses to study the effectiveness of our explanation under different settings, including multi-task learning and transfer learning. Our code and data are available at https://github.com/szzexpoi/rex.
Tree-structured multi-task architectures have been employed to jointly tackle multiple vision tasks in the context of multi-task learning (MTL). The major challenge is to determine where to branch out for each task given a backbone model to optimize for both task accuracy and computation efficiency. To address the challenge, this paper proposes a recommender that, given a set of tasks and a convolutional neural network-based backbone model, automatically suggests tree-structured multi-task architectures that could achieve a high task performance while meeting a user-specified computation budget without performing model training. Extensive evaluations on popular MTL benchmarks show that the recommended architectures could achieve competitive task accuracy and computation efficiency compared with state-of-the-art MTL methods. Our tree-structured multi-task model recommender is open-sourced and available at https://github.com/zhanglijun95/TreeMTL.
Multi-task learning for molecular property prediction is becoming increasingly important in drug discovery. However, in contrast to other domains, the performance of multi-task learning in drug discovery is still not satisfying as the number of labeled data for each task is too limited, which calls for additional data to complement the data scarcity. In this paper, we study multi-task learning for molecular property prediction in a novel setting, where a relation graph between tasks is available. We first construct a dataset (ChEMBL-STRING) including around 400 tasks as well as a task relation graph. Then to better utilize such relation graph, we propose a method called SGNN-EBM to systematically investigate the structured task modeling from two perspectives. (1) In the \emph{latent} space, we model the task representations by applying a state graph neural network (SGNN) on the relation graph. (2) In the \emph{output} space, we employ structured prediction with the energy-based model (EBM), which can be efficiently trained through noise-contrastive estimation (NCE) approach. Empirical results justify the effectiveness of SGNN-EBM. Code is available on https://github.com/chao1224/SGNN-EBM.
Deep learning-based underwater object detection (UOD) remains a major challenge due to the degraded visibility and difficulty to obtain sufficient underwater object images captured from various perspectives for training. To address these issues, this paper presents a high-precision UOD based on joint self-supervised deblurring and improved spatial transformer network. A self-supervised deblurring subnetwork is introduced into the designed multi-task learning aided object detection architecture to force the shared feature extraction module to output clean features for detection subnetwork. Aiming at alleviating the limitation of insufficient photos from different perspectives, an improved spatial transformer network is designed based on perspective transformation, adaptively enriching image features within the network. The experimental results show that the proposed UOD approach achieved 47.9 mAP in URPC2017 and 70.3 mAP in URPC2018, outperforming many state-of-the-art UOD methods and indicating the designed method is more suitable for UOD.
This paper proposes a high-precision self-supervised monocular VO, which is specifically designed for navigation in foggy weather. A cycled generative adversarial network is designed to obtain high-quality self-supervised loss via forcing the forward and backward half-cycle to output consistent estimation. Moreover, gradient-based loss and perceptual loss are introduced to eliminate the interference of complex photometric change on self-supervised loss in foggy weather. To solve the ill-posed problem of depth estimation, a self-supervised multi-task learning aided depth estimation module is designed based on the strong correlation between the depth estimation and transmission map calculation of hazy images in foggy weather. The experimental results on the synthetic foggy KITTI dataset show that the proposed self-supervised monocular VO performs better in depth and pose estimation than other state-of-the-art monocular VO in the literature, indicating the designed method is more suitable for foggy weather.
Building robust and general dialogue models for spoken conversations is challenging due to the gap in distributions of spoken and written data. This paper presents our approach to build generalized models for the Knowledge-grounded Task-oriented Dialogue Modeling on Spoken Conversations Challenge of DSTC-10. In order to mitigate the discrepancies between spoken and written text, we mainly employ extensive data augmentation strategies on written data, including artificial error injection and round-trip text-speech transformation. To train robust models for spoken conversations, we improve pre-trained language models, and apply ensemble algorithms for each sub-task. Typically, for the detection task, we fine-tune \roberta and ELECTRA, and run an error-fixing ensemble algorithm. For the selection task, we adopt a two-stage framework that consists of entity tracking and knowledge ranking, and propose a multi-task learning method to learn multi-level semantic information by domain classification and entity selection. For the generation task, we adopt a cross-validation data process to improve pre-trained generative language models, followed by a consensus decoding algorithm, which can add arbitrary features like relative \rouge metric, and tune associated feature weights toward \bleu directly. Our approach ranks third on the objective evaluation and second on the final official human evaluation.
The workflow of pretraining and fine-tuning has emerged as a popular paradigm for solving various NLP and V&L (Vision-and-Language) downstream tasks. With the capacity of pretrained models growing rapidly, how to perform parameter-efficient fine-tuning has become fairly important for quick transfer learning and deployment. In this paper, we design a novel unified parameter-efficient transfer learning framework that works effectively on both pure language and V&L tasks. In particular, we use a shared hypernetwork that takes trainable hyper-embeddings as input, and outputs weights for fine-tuning different small modules in a pretrained language model, such as tuning the parameters inserted into multi-head attention blocks (i.e., prefix-tuning) and feed-forward blocks (i.e., adapter-tuning). We define a set of embeddings (e.g., layer, block, task and visual embeddings) as the key components to calculate hyper-embeddings, which thus can support both pure language and V&L tasks. Our proposed framework adds fewer trainable parameters in multi-task learning while achieving superior performances and transfer ability compared to state-of-the-art methods. Empirical results on the GLUE benchmark and multiple V&L tasks confirm the effectiveness of our framework on both textual and visual modalities.
With the advancement of Deep Neural Networks (DNN) and large amounts of sensor data from Internet of Things (IoT) systems, the research community has worked to reduce the computational and resource demands of DNN to compute on low-resourced microcontrollers (MCUs). However, most of the current work in embedded deep learning focuses on solving a single task efficiently, while the multi-tasking nature and applications of IoT devices demand systems that can handle a diverse range of tasks (activity, voice, and context recognition) with input from a variety of sensors, simultaneously. In this paper, we propose YONO, a product quantization (PQ) based approach that compresses multiple heterogeneous models and enables in-memory model execution and switching for dissimilar multi-task learning on MCUs. We first adopt PQ to learn codebooks that store weights of different models. Also, we propose a novel network optimization and heuristics to maximize the compression rate and minimize the accuracy loss. Then, we develop an online component of YONO for efficient model execution and switching between multiple tasks on an MCU at run time without relying on an external storage device. YONO shows remarkable performance as it can compress multiple heterogeneous models with negligible or no loss of accuracy up to 12.37$\times$. Besides, YONO's online component enables an efficient execution (latency of 16-159 ms per operation) and reduces model loading/switching latency and energy consumption by 93.3-94.5% and 93.9-95.0%, respectively, compared to external storage access. Interestingly, YONO can compress various architectures trained with datasets that were not shown during YONO's offline codebook learning phase showing the generalizability of our method. To summarize, YONO shows great potential and opens further doors to enable multi-task learning systems on extremely resource-constrained devices.
This work presents Spacecraft Pose Network v2 (SPNv2), a Convolutional Neural Network (CNN) for pose estimation of noncooperative spacecraft across domain gap. SPNv2 is a multi-scale, multi-task CNN which consists of a shared multi-scale feature encoder and multiple prediction heads that perform different tasks on a shared feature output. These tasks are all related to detection and pose estimation of a target spacecraft from an image, such as prediction of pre-defined satellite keypoints, direct pose regression, and binary segmentation of the satellite foreground. It is shown that by jointly training on different yet related tasks with extensive data augmentations on synthetic images only, the shared encoder learns features that are common across image domains that have fundamentally different visual characteristics compared to synthetic images. This work also introduces Online Domain Refinement (ODR) which refines the parameters of the normalization layers of SPNv2 on the target domain images online at deployment. Specifically, ODR performs self-supervised entropy minimization of the predicted satellite foreground, thereby improving the CNN's performance on the target domain images without their pose labels and with minimal computational efforts. The GitHub repository for SPNv2 is available at https://github.com/tpark94/spnv2.
The study of multi-task learning has drawn great attention from the community. Despite the remarkable progress, the challenge of optimally learning different tasks simultaneously remains to be explored. Previous works attempt to modify the gradients from different tasks. Yet these methods give a subjective assumption of the relationship between tasks, and the modified gradient may be less accurate. In this paper, we introduce Stochastic Task Allocation~(STA), a mechanism that addresses this issue by a task allocation approach, in which each sample is randomly allocated a subset of tasks. For further progress, we propose Interleaved Stochastic Task Allocation~(ISTA) to iteratively allocate all tasks to each example during several consecutive iterations. We evaluate STA and ISTA on various datasets and applications: NYUv2, Cityscapes, and COCO for scene understanding and instance segmentation. Our experiments show both STA and ISTA outperform current state-of-the-art methods. The code will be available.
Prevailing deep models are single-purpose and overspecialize at individual tasks. However, when being extended to new tasks, they typically forget previously learned skills and learn from scratch. We address this issue by introducing SkillNet-NLU, a general-purpose model that stitches together existing skills to learn new tasks more effectively. The key feature of our approach is that it is sparsely activated guided by predefined skills. Different from traditional dense models that always activate all the model parameters, SkillNet-NLU only activates parts of the model parameters whose skills are relevant to the target task. When learning for a new task, our approach precisely activates required skills and also provides an option to add new skills. We evaluate on natural language understandings tasks and have the following findings. First, with only one model checkpoint, SkillNet-NLU performs better than task-specific fine-tuning and two multi-task learning baselines (i.e., dense model and Mixture-of-Experts model) on six tasks. Second, sparsely activated pre-training further improves the overall performance. Third, SkillNet-NLU significantly outperforms baseline systems when being extended to new tasks.
Active speaker detection and speech enhancement have become two increasingly attractive topics in audio-visual scenario understanding. According to their respective characteristics, the scheme of independently designed architecture has been widely used in correspondence to each single task. This may lead to the representation learned by the model being task-specific, and inevitably result in the lack of generalization ability of the feature based on multi-modal modeling. More recent studies have shown that establishing cross-modal relationship between auditory and visual stream is a promising solution for the challenge of audio-visual multi-task learning. Therefore, as a motivation to bridge the multi-modal associations in audio-visual tasks, a unified framework is proposed to achieve target speaker detection and speech enhancement with joint learning of audio-visual modeling in this study.
We consider a generic min-max multi-objective bilevel optimization problem with applications in robust machine learning such as representation learning and hyperparameter optimization. We design MORBiT, a novel single-loop gradient descent-ascent bilevel optimization algorithm, to solve the generic problem and present a novel analysis showing that MORBiT converges to the first-order stationary point at a rate of $\widetilde{\mathcal{O}}(n^{1/2} K^{-2/5})$ for a class of weakly convex problems with $n$ objectives upon $K$ iterations of the algorithm. Our analysis utilizes novel results to handle the non-smooth min-max multi-objective setup and to obtain a sublinear dependence in the number of objectives $n$. Experimental results on robust representation learning and robust hyperparameter optimization showcase (i) the advantages of considering the min-max multi-objective setup, and (ii) convergence properties of the proposed MORBiT. Our code is at https://github.com/minimario/MORBiT.