Instruction tuning of language models has demonstrated the ability to enhance model generalization to unseen tasks via in-context learning using a few examples. However, typical supervised learning still requires a plethora of downstream training data for finetuning. Often in real-world situations, there is a scarcity of data available for finetuning, falling somewhere between few shot inference and fully supervised finetuning. In this work, we demonstrate the sample efficiency of instruction tuned models over various tasks by estimating the minimal downstream training data required by them to perform transfer learning and match the performance of state-of-the-art (SOTA) supervised models. We conduct experiments on 119 tasks from Super Natural Instructions (SuperNI) in both the single task learning (STL) and multi task learning (MTL) settings. Our findings reveal that, in the STL setting, instruction tuned models equipped with 25% of the downstream train data surpass the SOTA performance on the downstream tasks. In the MTL setting, an instruction tuned model trained on only 6% of downstream training data achieve SOTA, while using 100% of the training data results in a 3.69% points improvement (ROUGE-L 74.68) over the previous SOTA. We conduct an analysis on T5 vs Tk-Instruct by developing several baselines to demonstrate that instruction tuning aids in increasing both sample efficiency and transfer learning. Additionally, we observe a consistent ~4% performance increase in both settings when pre-finetuning is performed with instructions. Finally, we conduct a categorical study and find that contrary to previous results, tasks in the question rewriting and title generation categories suffer from instruction tuning.
指令优化语言模型已经证明了通过使用几个例子来提高模型对未完成任务的泛化能力的能力。然而,典型的监督学习仍然需要大量的后续训练数据来进行微调。通常,在现实世界的情况下,微调数据的资源非常有限,处于几个Shot推断和完全监督微调之间的中间位置。在本文中,我们使用Super Natural Instructions(超级指令)中的119个任务进行了实验,同时在单任务学习和多任务学习环境中进行了测试。我们的发现表明,在STL环境中,指令优化模型所拥有的25%的后续训练数据超过了后续任务的性能表现(与SOTA相比)。在MTL环境中,只使用后续训练数据训练的指令优化模型达到了SOTA表现,而使用全部训练数据则导致了3.69%点的进步(ROUGE-L 74.68)高于之前的SOTA表现。我们对T5和Tk-Instruct进行了分析,以开发多个基准来表明指令优化有助于增加样本效率和迁移学习。此外,我们在两个环境中观察到一致的 ~4%的性能提升,在执行预微调指令之前。最后,我们进行了分类研究,并发现与之前的结果相反,问题改写和标题生成任务中的任务受到指令优化的影响。
https://arxiv.org/abs/2306.05539
3D reconstruction is a useful tool for surgical planning and guidance. However, the lack of available medical data stunts research and development in this field, as supervised deep learning methods for accurate disparity estimation rely heavily on large datasets containing ground truth information. Alternative approaches to supervision have been explored, such as self-supervision, which can reduce or remove entirely the need for ground truth. However, no proposed alternatives have demonstrated performance capabilities close to what would be expected from a supervised setup. This work aims to alleviate this issue. In this paper, we investigate the learning of structured light projections to enhance the development of direct disparity estimation networks. We show for the first time that it is possible to accurately learn the projection of structured light on a scene, implicitly learning disparity. Secondly, we \textcolor{black}{explore the use of a multi task learning (MTL) framework for the joint training of structured light and disparity. We present results which show that MTL with structured light improves disparity training; without increasing the number of model parameters. Our MTL setup outperformed the single task learning (STL) network in every validation test. Notably, in the medical generalisation test, the STL error was 1.4 times worse than that of the best MTL performance. The benefit of using MTL is emphasised when the training data is limited.} A dataset containing stereoscopic images, disparity maps and structured light projections on medical phantoms and ex vivo tissue was created for evaluation together with virtual scenes. This dataset will be made publicly available in the future.
https://arxiv.org/abs/2301.08140
Perceiving the surrounding environment is essential for enabling autonomous or assisted driving functionalities. Common tasks in this domain include detecting road users, as well as determining lane boundaries and classifying driving conditions. Over the last few years, a large variety of powerful Deep Learning models have been proposed to address individual tasks of camera-based automotive perception with astonishing performances. However, the limited capabilities of in-vehicle embedded computing platforms cannot cope with the computational effort required to run a heavy model for each individual task. In this work, we present CERBERUS (CEnteR Based End-to-end peRception Using a Single model), a lightweight model that leverages a multitask-learning approach to enable the execution of multiple perception tasks at the cost of a single inference. The code will be made publicly available at this https URL
https://arxiv.org/abs/2210.00756
The account of mitotic cells is a key feature in tumor diagnosis. However, due to the variability of mitotic cell morphology, it is a highly challenging task to detect mitotic cells in tumor tissues. At the same time, although advanced deep learning method have achieved great success in cell detection, the performance is often unsatisfactory when tested data from another domain (i.e. the different tumor types and different scanners). Therefore, it is necessary to develop algorithms for detecting mitotic cells with robustness in domain shifts scenarios. Our work further proposes a foreground detection and tumor classification task based on the baseline(Retinanet), and utilizes data augmentation to improve the domain generalization performance of our model. We achieve the state-of-the-art performance (F1 score: 0.5809) on the challenging premilary test dataset.
https://arxiv.org/abs/2208.12657
Massively Multilingual Transformer based Language Models have been observed to be surprisingly effective on zero-shot transfer across languages, though the performance varies from language to language depending on the pivot language(s) used for fine-tuning. In this work, we build upon some of the existing techniques for predicting the zero-shot performance on a task, by modeling it as a multi-task learning problem. We jointly train predictive models for different tasks which helps us build more accurate predictors for tasks where we have test data in very few languages to measure the actual performance of the model. Our approach also lends us the ability to perform a much more robust feature selection and identify a common set of features that influence zero-shot performance across a variety of tasks.
https://arxiv.org/abs/2205.06130
Due to the collection of big data and the development of deep learning, research to predict human emotions in the wild is being actively conducted. We designed a multi-task model using ABAW dataset to predict valence-arousal, expression, and action unit through audio data and face images at in real world. We trained model from the incomplete label by applying the knowledge distillation technique. The teacher model was trained as a supervised learning method, and the student model was trained by using the output of the teacher model as a soft label. As a result we achieved 2.40 in Multi Task Learning task validation dataset.
https://arxiv.org/abs/2203.13072
Retail item data contains many different forms of text like the title of an item, the description of an item, item name and reviews. It is of interest to identify the item name in the other forms of text using a named entity tagger. However, the title of an item and its description are syntactically different (but semantically similar) in that the title is not necessarily a well formed sentence while the description is made up of well formed sentences. In this work, we use a triplet loss to contrast the embeddings of the item title with the description to establish a proof of concept. We find that using the triplet loss in a multi-task NER algorithm improves both the precision and recall by a small percentage. While the improvement is small, we think it is a step in the right direction of using various forms of text in a multi-task algorithm. In addition to precision and recall, the multi task triplet loss method is also found to significantly improve the exact match accuracy i.e. the accuracy of tagging the entire set of tokens in the text with correct tags.
https://arxiv.org/abs/2109.13736
Dynamic balancing under uncertain disturbances is important for a humanoid robot, which requires a good capability of coordinating the entire body redundancy to execute multi tasks. Whole-body control (WBC) based on hierarchical optimization has been generally accepted and utilized in torque-controlled robots. A good hierarchy is the prerequisite for WBC and can be predefined according to prior knowledge. However, the real-time computation would be problematic in the physical applications considering the computational complexity of WBC. For robots with proprioceptive actuation, the joint friction in gear reducer would also degrade the torque tracking performance. In our paper, a reasonable hierarchy of tasks and constraints is first customized for robot dynamic balancing. Then a real-time WBC is implemented via a computationally efficient WBC software. Such a method is solved on a modular master control system UBTMaster characterized by the real-time communication and powerful computing capability. After the joint friction being well covered by the model identification, extensive experiments on various balancing scenarios are conducted on a humanoid Walker3 with proprioceptive actuation. The robot shows an outstanding balance performance even under external impulses as well as the two feet of the robot suffering the inclination and shift disturbances independently. The results demonstrate that with the strict hierarchy, real-time computation and joint friction being handled carefully, the robot with proprioceptive actuation can manage the dynamic physical interactions with the unstructured environments well.
https://arxiv.org/abs/2108.03826
The progress in Computer Aided Diagnosis (CADx) of Wireless Capsule Endoscopy (WCE) is thwarted by the lack of data. The inadequacy in richly representative healthy and abnormal conditions results in isolated analyses of pathologies, that can not handle realistic multi-pathology scenarios. In this work, we explore how to learn more for free, from limited data through solving a WCE multicentric, multi-pathology classification problem. Learning more implies to learning more than full supervision would allow with the same data. This is done by combining self supervision with full supervision, under multi task learning. Additionally, we draw inspiration from the Human Visual System (HVS) in designing self supervision tasks and investigate if seemingly ineffectual signals within the data itself can be exploited to gain performance, if so, which signals would be better than others. Further, we present our analysis of the high level features as a stepping stone towards more robust multi-pathology CADx in WCE.
https://arxiv.org/abs/2106.16162
The promising performance of Deep Neural Networks (DNNs) in text classification, has attracted researchers to use them for fraud review detection. However, the lack of trusted labeled data has limited the performance of the current solutions in detecting fraud reviews. The Generative Adversarial Network (GAN) as a semi-supervised method has demonstrated to be effective for data augmentation purposes. The state-of-the-art solutions utilize GANs to overcome the data scarcity problem. However, they fail to incorporate the behavioral clues in fraud generation. Additionally, state-of-the-art approaches overlook the possible bot-generated reviews in the dataset. Finally, they also suffer from a common limitation in scalability and stability of the GAN, slowing down the training procedure. In this work, we propose ScoreGAN for fraud review detection that makes use of both review text and review rating scores in the generation and detection process. Scores are incorporated through Information Gain Maximization (IGM) into the loss function for three reasons. One is to generate score-correlated reviews based on the scores given to the generator. Second, the generated reviews are employed to train the discriminator, so the discriminator can correctly label the possible bot-generated reviews through joint representations learned from the concatenation of GLobal Vector for Word representation (GLoVe) extracted from the text and the score. Finally, it can be used to improve the stability and scalability of the GAN. Results show that the proposed framework outperformed the existing state-of-the-art framework, namely FakeGAN, in terms of AP by 7\%, and 5\% on the Yelp and TripAdvisor datasets, respectively.
https://arxiv.org/abs/2006.06561
We are interested in learning models of non-stationary environments, which can be framed as a multi-task learning problem. Model-free reinforcement learning algorithms can achieve good asymptotic performance in multi-task learning at a cost of extensive sampling, due to their approach, which requires learning from scratch. While model-based approaches are among the most data efficient learning algorithms, they still struggle with complex tasks and model uncertainties. Meta-reinforcement learning addresses the efficiency and generalization challenges on multi task learning by quickly leveraging the meta-prior policy for a new task. In this paper, we propose a meta-reinforcement learning approach to learn the dynamic model of a non-stationary environment to be used for meta-policy optimization later. Due to the sample efficiency of model-based learning methods, we are able to simultaneously train both the meta-model of the non-stationary environment and the meta-policy until dynamic model convergence. Then, the meta-learned dynamic model of the environment will generate simulated data for meta-policy optimization. Our experiment demonstrates that our proposed method can meta-learn the policy in a non-stationary environment with the data efficiency of model-based learning approaches while achieving the high asymptotic performance of model-free meta-reinforcement learning.
https://arxiv.org/abs/2011.10714
Deep speaker embeddings have become the leading method for encoding speaker identity in speaker recognition tasks. The embedding space should ideally capture the variations between all possible speakers, encoding the multiple aspects that make up speaker identity. In this work, utilizing speaker age as an auxiliary variable in US Supreme Court recordings and speaker nationality with VoxCeleb, we show that by leveraging additional speaker attribute information in a multi task learning setting, deep speaker embedding performance can be increased for verification and diarization tasks, achieving a relative improvement of 17.8% in DER and 8.9% in EER for Supreme Court audio compared to omitting the auxiliary task. Experimental code has been made publicly available.
https://arxiv.org/abs/2010.14269
Deep Neural Networks (DNN) could forget the knowledge about earlier tasks when learning new tasks, and this is known as \textit{catastrophic forgetting}. While recent continual learning methods are capable of alleviating the catastrophic problem on toy-sized datasets, some issues still remain to be tackled when applying them in real-world problems. Recently, the fast mask-based learning method (e.g. piggyback \cite{mallya2018piggyback}) is proposed to address these issues by learning only a binary element-wise mask in a fast manner, while keeping the backbone model fixed. However, the binary mask has limited modeling capacity for new tasks. A more recent work \cite{hung2019compacting} proposes a compress-grow-based method (CPG) to achieve better accuracy for new tasks by partially training backbone model, but with order-higher training cost, which makes it infeasible to be deployed into popular state-of-the-art edge-/mobile-learning. The primary goal of this work is to simultaneously achieve fast and high-accuracy multi task adaption in continual learning setting. Thus motivated, we propose a new training method called \textit{kernel-wise Soft Mask} (KSM), which learns a kernel-wise hybrid binary and real-value soft mask for each task, while using the same backbone model. Such a soft mask can be viewed as a superposition of a binary mask and a properly scaled real-value tensor, which offers a richer representation capability without low-level kernel support to meet the objective of low hardware overhead. We validate KSM on multiple benchmark datasets against recent state-of-the-art methods (e.g. Piggyback, Packnet, CPG, etc.), which shows good improvement in both accuracy and training cost.
https://arxiv.org/abs/2009.05668
The attributes of object contours has great significance for instance segmentation task. However, most of the current popular deep neural networks do not pay much attention to the target edge information. Inspired by the human annotation process when making instance segmentation datasets, in this paper, we propose Mask Point RCNN aiming at promoting the neural networks attention to the target edge information, which can heighten the information propagates between multiple tasks by using different attributes features. Specifically, we present an auxiliary task to Mask RCNN, including utilizing keypoint detection technology to construct the target edge contour, and enhancing the sensitivity of the network to the object edge through multi task learning and feature fusion. These improvements are easy to implement and have a small amount of additional computing overhead. By extensive evaluations on the Cityscapes dataset, the results show that our approach outperforms vanilla Mask RCNN by 5.4 on the validation subset and 5.0 on the test subset.
https://arxiv.org/abs/2008.00460
Morphological information is important for many sequence labeling tasks in Natural Language Processing (NLP). Yet, existing approaches rely heavily on manual annotations or external software to capture this information. In this study, we propose using subword contextual embeddings to capture the morphological information for languages with rich morphology. In addition, we incorporate these embeddings in a hierarchical multi-task setting which is not employed before, to the best of our knowledge. Evaluated on Dependency Parsing (DEP) and Named Entity Recognition (NER) tasks, which are shown to benefit greatly from morphological information, our final model outperforms previous state-of-the-art models on both tasks for the Turkish language. Besides, we show a net improvement of 18.86% and 4.61% F-1 over the previously proposed multi-task learner in the same setting for the DEP and the NER tasks, respectively. Empirical results for five different MTL settings show that incorporating subword contextual embeddings brings significant improvements for both tasks. In addition, we observed that multi-task learning consistently improves the performance of the DEP component.
https://arxiv.org/abs/2004.12247
We propose a multi task learning-based neural model for bridging reference resolution tackling two key challenges faced by bridging reference resolution. The first challenge is the lack of large corpora annotated with bridging references. To address this, we use multi-task learning to help bridging reference resolution with coreference resolution. We show that substantial improvements of up to 8 p.p. can be achieved on full bridging resolution with this architecture. The second challenge is the different definitions of bridging used in different corpora, meaning that hand-coded systems or systems using special features designed for one corpus do not work well with other corpora. Our neural model only uses a small number of corpus independent features, thus can be applied easily to different corpora. Evaluations with very different bridging corpora (ARRAU, ISNOTES, BASHI and SCICORP) suggest that our architecture works equally well on all corpora, and achieves the SoTA results on full bridging resolution for all corpora, outperforming the best reported results by up to 34.9 percentage points.
https://arxiv.org/abs/2003.03666
We propose a new semi-supervised learning method on face-related tasks based on Multi-Task Learning (MTL) and data distillation. The proposed method exploits multiple datasets with different labels for different-but-related tasks such as simultaneous age, gender, race, facial expression estimation. Specifically, when there are only a few well-labeled data for a specific task among the multiple related ones, we exploit the labels of other related tasks in different domains. Our approach is composed of (1) a new MTL method which can deal with weakly labeled datasets and perform several tasks simultaneously, and (2) an MTL-based data distillation framework which enables network generalization for the training and test data from different domains. Experiments show that the proposed multi-task system performs each task better than the baseline single task. It is also demonstrated that using different domain datasets along with the main dataset can enhance network generalization and overcome the domain differences between datasets. Also, comparing data distillation both on the baseline and MTL framework, the latter shows more accurate predictions on unlabeled data from different domains. Furthermore, by proposing a new learning-rate optimization method, our proposed network is able to dynamically tune its learning rate.
https://arxiv.org/abs/1907.03402
This paper proposes a method to gain extra supervision via multi-task learning for multi-modal video question answering. Multi-modal video question answering is an important task that aims at the joint understanding of vision and language. However, establishing large scale dataset for multi-modal video question answering is expensive and the existing benchmarks are relatively small to provide sufficient supervision. To overcome this challenge, this paper proposes a multi-task learning method which is composed of three main components: (1) multi-modal video question answering network that answers the question based on the both video and subtitle feature, (2) temporal retrieval network that predicts the time in the video clip where the question was generated from and (3) modality alignment network that solves metric learning problem to find correct association of video and subtitle modalities. By simultaneously solving related auxiliary tasks with hierarchically shared intermediate layers, the extra synergistic supervisions are provided. Motivated by curriculum learning, multi task ratio scheduling is proposed to learn easier task earlier to set inductive bias at the beginning of the training. The experiments on publicly available dataset TVQA shows state-of-the-art results, and ablation studies are conducted to prove the statistical validity.
本文提出了一种通过多任务学习获得额外监督的多模式视频问答方法。多模态视频问答是一项旨在实现视觉和语言共同理解的重要任务。然而,建立大规模的多模式视频问答数据集代价高昂,现有的基准相对较小,无法提供足够的监督。为了克服这一挑战,本文提出了一种多任务学习方法,该方法由三个主要部分组成:(1)基于视频和字幕特征的多模式视频问答网络;(2)时间检索网络,预测视频剪辑中生成问题的时间。OM和(3)模态对齐网络,解决了度量学习问题,以找到视频和副标题模态的正确关联。通过同时解决具有层次共享中间层的相关辅助任务,提供了额外的协同监控。在课程学习的激励下,提出了多任务比调度的方法,以便在培训开始时提前学习更容易的任务,从而设置归纳偏差。在公共可用数据集TVCA上进行的实验显示了最新的结果,并进行了消融研究以证明统计有效性。
https://arxiv.org/abs/1905.13540
Many prediction tasks, especially in computer vision, are often inherently ambiguous. For example, the output of semantic segmentation may depend on the scale one is looking at, and image saliency or video summarization is often user or context dependent. Arguably, in such scenarios, exploiting instance specific evidence, such as scale or user context, can help resolve the underlying ambiguity leading to the improved predictions. While existing literature has considered incorporating such evidence in classical models such as probabilistic graphical models (PGMs), there is limited (or no) prior work looking at this problem in the context of deep neural network (DNN) models. In this paper, we present a generic multi task learning (MTL) based framework which handles the evidence as the output of one or more secondary tasks, while modeling the original problem as the primary task of interest. Our training phase is identical to the one used by standard MTL architectures. During prediction, we back-propagate the loss on secondary task(s) such that network weights are re-adjusted to match the evidence. An early stopping or two norm based regularizer ensures weights do not deviate significantly from the ones learned originally. Implementation in two specific scenarios (a) predicting semantic segmentation given the image level tags (b) predicting instance level segmentation given the text description of the image, clearly demonstrates the effectiveness of our proposed approach.
https://arxiv.org/abs/1811.09796
Morphological analysis is an important first step in downstream tasks like machine translation and dependency parsing of morphologically rich languages (MRLs) such as those belonging to Indo-Aryan and Dravidian families. However, the ambiguities introduced by the recombination of morphemes constructing several possible inflections for a word makes the prediction of syntactic traits a notoriously complicated task for MRLs. We propose a character-level neural morphological analyzer, the Multi Task Deep Morphological analyzer (MT-DMA), based on multitask learning of word-level tag markers for Hindi. In order to show the portability of our system to other related languages, we present results on Urdu too. MT-DMA predicts the complete set of morphological tags for words of Indo-Aryan languages: Parts-of-speech (POS), Gender (G), Number (N), Person (P), Case (C), Tense-Aspect-Modality (TAM) marker as well as the Lemma (L) by jointly learning all these in a single end-to-end framework. We show the effectiveness of training of such deep neural networks by the simultaneous optimization of multiple loss functions and sharing of initial parameters for context-aware morphological analysis. Our model outperforms the state-of-art analyzers for Hindi and Urdu. Exploring the use of a set of character-level features in phonological space optimized for each tag through a multi-objective genetic algorithm, coupled with effective training strategies, our model establishes a new state-of-the-art accuracy score upon all seven of the tasks for both the languages. MT-DMA is publicly accessible to be used at <a href="http://35.154.251.44/.">this http URL</a>
https://arxiv.org/abs/1811.08619