Though modern neural networks have achieved impressive performance in both vision and language tasks, we know little about the functions that they implement. One possibility is that neural networks implicitly break down complex tasks into subroutines, implement modular solutions to these subroutines, and compose them into an overall solution to a task - a property we term structural compositionality. Another possibility is that they may simply learn to match new inputs to learned templates, eliding task decomposition entirely. Here, we leverage model pruning techniques to investigate this question in both vision and language across a variety of architectures, tasks, and pretraining regimens. Our results demonstrate that models often implement solutions to subroutines via modular subnetworks, which can be ablated while maintaining the functionality of other subnetworks. This suggests that neural networks may be able to learn compositionality, obviating the need for specialized symbolic mechanisms.
相關內容
With the rapid development of Quantum Machine Learning, quantum neural networks (QNN) have experienced great advancement in the past few years, harnessing the advantages of quantum computing to significantly speed up classical machine learning tasks. Despite their increasing popularity, the quantum neural network is quite counter-intuitive and difficult to understand, due to their unique quantum-specific layers (e.g., data encoding and measurement) in their architecture. It prevents QNN users and researchers from effectively understanding its inner workings and exploring the model training status. To fill the research gap, we propose VIOLET, a novel visual analytics approach to improve the explainability of quantum neural networks. Guided by the design requirements distilled from the interviews with domain experts and the literature survey, we developed three visualization views: the Encoder View unveils the process of converting classical input data into quantum states, the Ansatz View reveals the temporal evolution of quantum states in the training process, and the Feature View displays the features a QNN has learned after the training process. Two novel visual designs, i.e., satellite chart and augmented heatmap, are proposed to visually explain the variational parameters and quantum circuit measurements respectively. We evaluate VIOLET through two case studies and in-depth interviews with 12 domain experts. The results demonstrate the effectiveness and usability of VIOLET in helping QNN users and developers intuitively understand and explore quantum neural networks
In this survey, we examine algorithms for conducting credit assignment in artificial neural networks that are inspired or motivated by neurobiology. These processes are unified under one possible taxonomy, which is constructed based on how a learning algorithm answers a central question underpinning the mechanisms of synaptic plasticity in complex adaptive neuronal systems: where do the signals that drive the learning in individual elements of a network come from and how are they produced? In this unified treatment, we organize the ever-growing set of brain-inspired learning schemes into six general families and consider these in the context of backpropagation of errors and its known criticisms. The results of this review are meant to encourage future developments in neuro-mimetic systems and their constituent learning processes, wherein lies an important opportunity to build a strong bridge between machine learning, computational neuroscience, and cognitive science.
Purpose: One of the major reasons that totally implantable cochlear microphones are not readily available is the lack of good implantable microphones. An implantable microphone has the potential to provide a range of benefits over external microphones for cochlear implant users including the filtering ability of the outer ear, cosmetics, and usability in all situations. This paper presents results from experiments in human cadaveric ears of a piezofilm microphone concept under development as a possible component of a future implantable microphone system for use with cochlear implants. This microphone is referred to here as a drum microphone (DrumMic) that senses the robust and predictable motion of the umbo, the tip of the malleus. Methods: The performance was measured of five DrumMics inserted in four different human cadaveric temporal bones. Sensitivity, linearity, bandwidth, and equivalent input noise were measured during these experiments using a sound stimulus and measurement setup. Results: The sensitivity of the DrumMics was found to be tightly clustered across different microphones and ears despite differences in umbo and middle ear anatomy. The DrumMics were shown to behave linearly across a large dynamic range (46 dB SPL to 100 dB SPL) across a wide bandwidth (100 Hz to 8 kHz). The equivalent input noise (0.1-10 kHz) of the DrumMic and amplifier referenced to the ear canal was measured to be 54 dB SPL and estimated to be 46 dB SPL after accounting for the pressure gain of the outer ear. Conclusion: The results demonstrate that the DrumMic behaves robustly across ears and fabrication. The equivalent input noise performance was shown to approach that of commercial hearing aid microphones. To advance this demonstration of the DrumMic concept to a future prototype implantable in humans, work on encapsulation, biocompatibility, connectorization will be required.
Transformer is a promising neural network learner, and has achieved great success in various machine learning tasks. Thanks to the recent prevalence of multimodal applications and big data, Transformer-based multimodal learning has become a hot topic in AI research. This paper presents a comprehensive survey of Transformer techniques oriented at multimodal data. The main contents of this survey include: (1) a background of multimodal learning, Transformer ecosystem, and the multimodal big data era, (2) a theoretical review of Vanilla Transformer, Vision Transformer, and multimodal Transformers, from a geometrically topological perspective, (3) a review of multimodal Transformer applications, via two important paradigms, i.e., for multimodal pretraining and for specific multimodal tasks, (4) a summary of the common challenges and designs shared by the multimodal Transformer models and applications, and (5) a discussion of open problems and potential research directions for the community.
Transformers have achieved superior performances in many tasks in natural language processing and computer vision, which also intrigues great interests in the time series community. Among multiple advantages of transformers, the ability to capture long-range dependencies and interactions is especially attractive for time series modeling, leading to exciting progress in various time series applications. In this paper, we systematically review transformer schemes for time series modeling by highlighting their strengths as well as limitations through a new taxonomy to summarize existing time series transformers in two perspectives. From the perspective of network modifications, we summarize the adaptations of module level and architecture level of the time series transformers. From the perspective of applications, we categorize time series transformers based on common tasks including forecasting, anomaly detection, and classification. Empirically, we perform robust analysis, model size analysis, and seasonal-trend decomposition analysis to study how Transformers perform in time series. Finally, we discuss and suggest future directions to provide useful research guidance. To the best of our knowledge, this paper is the first work to comprehensively and systematically summarize the recent advances of Transformers for modeling time series data. We hope this survey will ignite further research interests in time series Transformers.
Deep neural networks (DNNs) have become a proven and indispensable machine learning tool. As a black-box model, it remains difficult to diagnose what aspects of the model's input drive the decisions of a DNN. In countless real-world domains, from legislation and law enforcement to healthcare, such diagnosis is essential to ensure that DNN decisions are driven by aspects appropriate in the context of its use. The development of methods and studies enabling the explanation of a DNN's decisions has thus blossomed into an active, broad area of research. A practitioner wanting to study explainable deep learning may be intimidated by the plethora of orthogonal directions the field has taken. This complexity is further exacerbated by competing definitions of what it means ``to explain'' the actions of a DNN and to evaluate an approach's ``ability to explain''. This article offers a field guide to explore the space of explainable deep learning aimed at those uninitiated in the field. The field guide: i) Introduces three simple dimensions defining the space of foundational methods that contribute to explainable deep learning, ii) discusses the evaluations for model explanations, iii) places explainability in the context of other related deep learning research areas, and iv) finally elaborates on user-oriented explanation designing and potential future directions on explainable deep learning. We hope the guide is used as an easy-to-digest starting point for those just embarking on research in this field.
It has been shown that deep neural networks are prone to overfitting on biased training data. Towards addressing this issue, meta-learning employs a meta model for correcting the training bias. Despite the promising performances, super slow training is currently the bottleneck in the meta learning approaches. In this paper, we introduce a novel Faster Meta Update Strategy (FaMUS) to replace the most expensive step in the meta gradient computation with a faster layer-wise approximation. We empirically find that FaMUS yields not only a reasonably accurate but also a low-variance approximation of the meta gradient. We conduct extensive experiments to verify the proposed method on two tasks. We show our method is able to save two-thirds of the training time while still maintaining the comparable or achieving even better generalization performance. In particular, our method achieves the state-of-the-art performance on both synthetic and realistic noisy labels, and obtains promising performance on long-tailed recognition on standard benchmarks.
Deep neural networks (DNNs) are successful in many computer vision tasks. However, the most accurate DNNs require millions of parameters and operations, making them energy, computation and memory intensive. This impedes the deployment of large DNNs in low-power devices with limited compute resources. Recent research improves DNN models by reducing the memory requirement, energy consumption, and number of operations without significantly decreasing the accuracy. This paper surveys the progress of low-power deep learning and computer vision, specifically in regards to inference, and discusses the methods for compacting and accelerating DNN models. The techniques can be divided into four major categories: (1) parameter quantization and pruning, (2) compressed convolutional filters and matrix factorization, (3) network architecture search, and (4) knowledge distillation. We analyze the accuracy, advantages, disadvantages, and potential solutions to the problems with the techniques in each category. We also discuss new evaluation metrics as a guideline for future research.
With the advent of deep neural networks, learning-based approaches for 3D reconstruction have gained popularity. However, unlike for images, in 3D there is no canonical representation which is both computationally and memory efficient yet allows for representing high-resolution geometry of arbitrary topology. Many of the state-of-the-art learning-based 3D reconstruction approaches can hence only represent very coarse 3D geometry or are limited to a restricted domain. In this paper, we propose occupancy networks, a new representation for learning-based 3D reconstruction methods. Occupancy networks implicitly represent the 3D surface as the continuous decision boundary of a deep neural network classifier. In contrast to existing approaches, our representation encodes a description of the 3D output at infinite resolution without excessive memory footprint. We validate that our representation can efficiently encode 3D structure and can be inferred from various kinds of input. Our experiments demonstrate competitive results, both qualitatively and quantitatively, for the challenging tasks of 3D reconstruction from single images, noisy point clouds and coarse discrete voxel grids. We believe that occupancy networks will become a useful tool in a wide variety of learning-based 3D tasks.
Convolutional networks (ConvNets) have achieved great successes in various challenging vision tasks. However, the performance of ConvNets would degrade when encountering the domain shift. The domain adaptation is more significant while challenging in the field of biomedical image analysis, where cross-modality data have largely different distributions. Given that annotating the medical data is especially expensive, the supervised transfer learning approaches are not quite optimal. In this paper, we propose an unsupervised domain adaptation framework with adversarial learning for cross-modality biomedical image segmentations. Specifically, our model is based on a dilated fully convolutional network for pixel-wise prediction. Moreover, we build a plug-and-play domain adaptation module (DAM) to map the target input to features which are aligned with source domain feature space. A domain critic module (DCM) is set up for discriminating the feature space of both domains. We optimize the DAM and DCM via an adversarial loss without using any target domain label. Our proposed method is validated by adapting a ConvNet trained with MRI images to unpaired CT data for cardiac structures segmentations, and achieved very promising results.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191