Positioning is an essential service for various applications and is expected to be integrated with existing communication infrastructures in 5G and 6G. Though current Wi-Fi and cellular base stations (BSs) can be used to support this integration, the resulting precision is unsatisfactory due to the lack of precise control of the wireless signals. Recently, BSs adopting reconfigurable holographic surfaces (RHSs) have been advocated for positioning as RHSs' large number of antenna elements enable generation of arbitrary and highly-focused signal beam patterns. However, existing designs face two major challenges: i) RHSs only have limited operating bandwidth, and ii) the positioning methods cannot adapt to the diverse environments encountered in practice. To overcome these challenges, we present HoloFed, a system providing high-precision environment-adaptive user positioning services by exploiting multi-band(MB)-RHS and federated learning (FL). For improving the positioning performance, a lower bound on the error variance is obtained and utilized for guiding MB-RHS's digital and analog beamforming design. For better adaptability while preserving privacy, an FL framework is proposed for users to collaboratively train a position estimator, where we exploit the transfer learning technique to handle the lack of position labels of the users. Moreover, a scheduling algorithm for the BS to select which users train the position estimator is designed, jointly considering the convergence and efficiency of FL. Our simulation results confirm that HoloFed achieves a 57% lower positioning error variance compared to a beam-scanning baseline and can effectively adapt to diverse environments.
The increasing ubiquity of language technology necessitates a shift towards considering cultural diversity in the machine learning realm, particularly for subjective tasks that rely heavily on cultural nuances, such as Offensive Language Detection (OLD). Current understanding underscores that these tasks are substantially influenced by cultural values, however, a notable gap exists in determining if cultural features can accurately predict the success of cross-cultural transfer learning for such subjective tasks. Addressing this, our study delves into the intersection of cultural features and transfer learning effectiveness. The findings reveal that cultural value surveys indeed possess a predictive power for cross-cultural transfer learning success in OLD tasks and that it can be further improved using offensive word distance. Based on these results, we advocate for the integration of cultural information into datasets. Additionally, we recommend leveraging data sources rich in cultural information, such as surveys, to enhance cultural adaptability. Our research signifies a step forward in the quest for more inclusive, culturally sensitive language technologies.
The forthcoming 6G-enabled Intelligent Transportation System (ITS) is set to redefine conventional transportation networks with advanced intelligent services and applications. These technologies, including edge computing, Machine Learning (ML), and network softwarization, pose stringent requirements for latency, energy efficiency, and user data security. Distributed Learning (DL), such as Federated Learning (FL), is essential to meet these demands by distributing the learning process at the network edge. However, traditional FL approaches often require substantial resources for satisfactory learning performance. In contrast, Transfer Learning (TL) and Split Learning (SL) have shown effectiveness in enhancing learning efficiency in resource-constrained wireless scenarios like ITS. Non-terrestrial Networks (NTNs) have recently acquired a central place in the 6G vision, especially for boosting the coverage, capacity, and resilience of traditional terrestrial facilities. Air-based NTN layers, such as High Altitude Platforms (HAPs), can have added advantages in terms of reduced transmission distances and flexible deployments and thus can be exploited to enable intelligent solutions for latency-critical vehicular scenarios. With this motivation, in this work, we introduce the concept of Federated Split Transfer Learning (FSTL) in joint air-ground networks for resource-constrained vehicular scenarios. Simulations carried out in vehicular scenarios validate the efficacy of FSTL on HAPs in NTN, demonstrating significant improvements in addressing the demands of ITS applications.
Owing to advancements in deep learning technology, Vision Transformers (ViTs) have demonstrated impressive performance in various computer vision tasks. Nonetheless, ViTs still face some challenges, such as high computational complexity and the absence of desirable inductive biases. To alleviate these issues, {the potential advantages of combining eagle vision with ViTs are explored. We summarize a Bi-Fovea Visual Interaction (BFVI) structure inspired by the unique physiological and visual characteristics of eagle eyes. A novel Bi-Fovea Self-Attention (BFSA) mechanism and Bi-Fovea Feedforward Network (BFFN) are proposed based on this structural design approach, which can be used to mimic the hierarchical and parallel information processing scheme of the biological visual cortex, enabling networks to learn feature representations of targets in a coarse-to-fine manner. Furthermore, a Bionic Eagle Vision (BEV) block is designed as the basic building unit based on the BFSA mechanism and BFFN. By stacking BEV blocks, a unified and efficient family of pyramid backbone networks called Eagle Vision Transformers (EViTs) is developed. Experimental results show that EViTs exhibit highly competitive performance in various computer vision tasks, such as image classification, object detection and semantic segmentation. Compared with other approaches, EViTs have significant advantages, especially in terms of performance and computational efficiency. Code is available at https://github.com/nkusyl/EViT
Dataset distillation methods have achieved remarkable success in distilling a large dataset into a small set of representative samples. However, they are not designed to produce a distilled dataset that can be effectively used for facilitating self-supervised pre-training. To this end, we propose a novel problem of distilling an unlabeled dataset into a set of small synthetic samples for efficient self-supervised learning (SSL). We first prove that a gradient of synthetic samples with respect to a SSL objective in naive bilevel optimization is \textit{biased} due to the randomness originating from data augmentations or masking. To address this issue, we propose to minimize the mean squared error (MSE) between a model's representations of the synthetic examples and their corresponding learnable target feature representations for the inner objective, which does not introduce any randomness. Our primary motivation is that the model obtained by the proposed inner optimization can mimic the \textit{self-supervised target model}. To achieve this, we also introduce the MSE between representations of the inner model and the self-supervised target model on the original full dataset for outer optimization. Lastly, assuming that a feature extractor is fixed, we only optimize a linear head on top of the feature extractor, which allows us to reduce the computational cost and obtain a closed-form solution of the head with kernel ridge regression. We empirically validate the effectiveness of our method on various applications involving transfer learning.
A physics-informed convolutional neural network is proposed to simulate two phase flow in porous media with time-varying well controls. While most of PICNNs in existing literatures worked on parameter-to-state mapping, our proposed network parameterizes the solution with time-varying controls to establish a control-to-state regression. Firstly, finite volume scheme is adopted to discretize flow equations and formulate loss function that respects mass conservation laws. Neumann boundary conditions are seamlessly incorporated into the semi-discretized equations so no additional loss term is needed. The network architecture comprises two parallel U-Net structures, with network inputs being well controls and outputs being the system states. To capture the time-dependent relationship between inputs and outputs, the network is well designed to mimic discretized state space equations. We train the network progressively for every timestep, enabling it to simultaneously predict oil pressure and water saturation at each timestep. After training the network for one timestep, we leverage transfer learning techniques to expedite the training process for subsequent timestep. The proposed model is used to simulate oil-water porous flow scenarios with varying reservoir gridblocks and aspects including computation efficiency and accuracy are compared against corresponding numerical approaches. The results underscore the potential of PICNN in effectively simulating systems with numerous grid blocks, as computation time does not scale with model dimensionality. We assess the temporal error using 10 different testing controls with variation in magnitude and another 10 with higher alternation frequency with proposed control-to-state architecture. Our observations suggest the need for a more robust and reliable model when dealing with controls that exhibit significant variations in magnitude or frequency.
Lung diseases such as COVID-19, tuberculosis (TB), and pneumonia continue to be serious global health concerns that affect millions of people worldwide. In medical practice, chest X-ray examinations have emerged as the norm for diagnosing diseases, particularly chest infections such as COVID-19. Paramedics and scientists are working intensively to create a reliable and precise approach for early-stage COVID-19 diagnosis in order to save lives. But with a variety of symptoms, medical diagnosis of these disorders poses special difficulties. It is essential to address their identification and timely diagnosis in order to successfully treat and prevent these illnesses. In this research, a multiclass classification approach using state-of-the-art methods for deep learning and image processing is proposed. This method takes into account the robustness and efficiency of the system in order to increase diagnostic precision of chest diseases. A comparison between a brand-new convolution neural network (CNN) and several transfer learning pre-trained models including VGG19, ResNet, DenseNet, EfficientNet, and InceptionNet is recommended. Publicly available and widely used research datasets like Shenzen, Montogomery, the multiclass Kaggle dataset and the NIH dataset were used to rigorously test the model. Recall, precision, F1-score, and Area Under Curve (AUC) score are used to evaluate and compare the performance of the proposed model. An AUC value of 0.95 for COVID-19, 0.99 for TB, and 0.98 for pneumonia is obtained using the proposed network. Recall and precision ratings of 0.95, 0.98, and 0.97, respectively, likewise met high standards.
Deep neural networks can yield good performance on various tasks but often require large amounts of data to train them. Meta-learning received considerable attention as one approach to improve the generalization of these networks from a limited amount of data. Whilst meta-learning techniques have been observed to be successful at this in various scenarios, recent results suggest that when evaluated on tasks from a different data distribution than the one used for training, a baseline that simply finetunes a pre-trained network may be more effective than more complicated meta-learning techniques such as MAML, which is one of the most popular meta-learning techniques. This is surprising as the learning behaviour of MAML mimics that of finetuning: both rely on re-using learned features. We investigate the observed performance differences between finetuning, MAML, and another meta-learning technique called Reptile, and show that MAML and Reptile specialize for fast adaptation in low-data regimes of similar data distribution as the one used for training. Our findings show that both the output layer and the noisy training conditions induced by data scarcity play important roles in facilitating this specialization for MAML. Lastly, we show that the pre-trained features as obtained by the finetuning baseline are more diverse and discriminative than those learned by MAML and Reptile. Due to this lack of diversity and distribution specialization, MAML and Reptile may fail to generalize to out-of-distribution tasks whereas finetuning can fall back on the diversity of the learned features.
Transfer learning is a crucial technique for handling a small amount of data that is potentially related to other abundant data. However, most of the existing methods are focused on classification tasks using images and language datasets. Therefore, in order to expand the transfer learning scheme to regression tasks, we propose a novel transfer technique based on differential geometry, namely the Geometrically Aligned Transfer Encoder (GATE). In this method, we interpret the latent vectors from the model to exist on a Riemannian curved manifold. We find a proper diffeomorphism between pairs of tasks to ensure that every arbitrary point maps to a locally flat coordinate in the overlapping region, allowing the transfer of knowledge from the source to the target data. This also serves as an effective regularizer for the model to behave in extrapolation regions. In this article, we demonstrate that GATE outperforms conventional methods and exhibits stable behavior in both the latent space and extrapolation regions for various molecular graph datasets.
Recent advances in general purpose pre-trained language models have shown great potential in commonsense reasoning. However, current works still perform poorly on standard commonsense reasoning benchmarks including the Com2Sense Dataset. We argue that this is due to a disconnect with current cutting-edge machine learning methods. In this work, we aim to bridge the gap by introducing current ML-based methods to improve general purpose pre-trained language models in the task of commonsense reasoning. Specifically, we experiment with and systematically evaluate methods including knowledge transfer, model ensemble, and introducing an additional pairwise contrastive objective. Our best model outperforms the strongest previous works by ~15\% absolute gains in Pairwise Accuracy and ~8.7\% absolute gains in Standard Accuracy.
Objective To develop soft prompt-based learning algorithms for large language models (LLMs), examine the shape of prompts, prompt-tuning using frozen/unfrozen LLMs, transfer learning, and few-shot learning abilities. Methods We developed a soft prompt-based LLM model and compared 4 training strategies including (1) fine-tuning without prompts; (2) hard-prompt with unfrozen LLMs; (3) soft-prompt with unfrozen LLMs; and (4) soft-prompt with frozen LLMs. We evaluated 7 pretrained LLMs using the 4 training strategies for clinical concept and relation extraction on two benchmark datasets. We evaluated the transfer learning ability of the prompt-based learning algorithms in a cross-institution setting. We also assessed the few-shot learning ability. Results and Conclusion When LLMs are unfrozen, GatorTron-3.9B with soft prompting achieves the best strict F1-scores of 0.9118 and 0.8604 for concept extraction, outperforming the traditional fine-tuning and hard prompt-based models by 0.6~3.1% and 1.2~2.9%, respectively; GatorTron-345M with soft prompting achieves the best F1-scores of 0.8332 and 0.7488 for end-to-end relation extraction, outperforming the other two models by 0.2~2% and 0.6~11.7%, respectively. When LLMs are frozen, small (i.e., 345 million parameters) LLMs have a big gap to be competitive with unfrozen models; scaling LLMs up to billions of parameters makes frozen LLMs competitive with unfrozen LLMs. For cross-institute evaluation, soft prompting with a frozen GatorTron-8.9B model achieved the best performance. This study demonstrates that (1) machines can learn soft prompts better than humans, (2) frozen LLMs have better few-shot learning ability and transfer learning ability to facilitate muti-institution applications, and (3) frozen LLMs require large models.
Brain tumor segmentation presents a formidable challenge in the field of Medical Image Segmentation. While deep-learning models have been useful, human expert segmentation remains the most accurate method. The recently released Segment Anything Model (SAM) has opened up the opportunity to apply foundation models to this difficult task. However, SAM was primarily trained on diverse natural images. This makes applying SAM to biomedical segmentation, such as brain tumors with less defined boundaries, challenging. In this paper, we enhanced SAM's mask decoder using transfer learning with the Decathlon brain tumor dataset. We developed three methods to encapsulate the four-dimensional data into three dimensions for SAM. An on-the-fly data augmentation approach has been used with a combination of rotations and elastic deformations to increase the size of the training dataset. Two key metrics: the Dice Similarity Coefficient (DSC) and the Hausdorff Distance 95th Percentile (HD95), have been applied to assess the performance of our segmentation models. These metrics provided valuable insights into the quality of the segmentation results. In our evaluation, we compared this improved model to two benchmarks: the pretrained SAM and the widely used model, nnUNetv2. We find that the improved SAM shows considerable improvement over the pretrained SAM, while nnUNetv2 outperformed the improved SAM in terms of overall segmentation accuracy. Nevertheless, the improved SAM demonstrated slightly more consistent results than nnUNetv2, especially on challenging cases that can lead to larger Hausdorff distances. In the future, more advanced techniques can be applied in order to further improve the performance of SAM on brain tumor segmentation.
Fog computing emerged as a promising paradigm to address the challenges of processing and managing data generated by the Internet of Things (IoT). Load balancing (LB) plays a crucial role in Fog computing environments to optimize the overall system performance. It requires efficient resource allocation to improve resource utilization, minimize latency, and enhance the quality of service for end-users. In this work, we improve the performance of privacy-aware Reinforcement Learning (RL) agents that optimize the execution delay of IoT applications by minimizing the waiting delay. To maintain privacy, these agents optimize the waiting delay by minimizing the change in the number of queued requests in the whole system, i.e., without explicitly observing the actual number of requests that are queued in each Fog node nor observing the compute resource capabilities of those nodes. Besides improving the performance of these agents, we propose in this paper a lifelong learning framework for these agents, where lightweight inference models are used during deployment to minimize action delay and only retrained in case of significant environmental changes. To improve the performance, minimize the training cost, and adapt the agents to those changes, we explore the application of Transfer Learning (TL). TL transfers the knowledge acquired from a source domain and applies it to a target domain, enabling the reuse of learned policies and experiences. TL can be also used to pre-train the agent in simulation before fine-tuning it in the real environment; this significantly reduces failure probability compared to learning from scratch in the real environment. To our knowledge, there are no existing efforts in the literature that use TL to address lifelong learning for RL-based Fog LB; this is one of the main obstacles in deploying RL LB solutions in Fog systems.
Time series has been left behind in the era of pre-training and transfer learning. While research in the fields of natural language processing and computer vision are enjoying progressively larger datasets to train massive models, the most popular time series datasets consist of only tens of thousands of time steps, limiting our ability to study the effectiveness of pre-training and scaling. Recent studies have also cast doubt on the need for expressive models and scale. To alleviate these issues, we introduce three large-scale time series forecasting datasets from the cloud operations (CloudOps) domain, the largest having billions of observations, enabling further study into pre-training and scaling of time series models. We build the empirical groundwork for studying pre-training and scaling of time series models and pave the way for future research by identifying a promising candidate architecture. We show that it is a strong zero-shot baseline and benefits from further scaling, both in model and dataset size. Accompanying these datasets and results is a suite of comprehensive benchmark results comparing classical and deep learning baselines to our pre-trained method - achieving a 27% reduction in error on the largest dataset. Code and datasets can be found https://github.com/SalesforceAIResearch/pretrain-time-series-cloudops.
Supporting ultra-high data rates and flexible reconfigurability, Terahertz (THz) mesh networks are attractive for next-generation wireless backhaul systems that empower the integrated access and backhaul (IAB). In THz mesh backhaul networks, the efficient cross-layer routing and long-term resource allocation is yet an open problem due to dynamic traffic demands as well as possible link failures caused by the high directivity and high non-line-of-sight (NLoS) path loss of THz spectrum. In addition, unpredictable data traffic and the mixed integer programming property with the NP-hard nature further challenge the effective routing and long-term resource allocation design. In this paper, a deep reinforcement learning (DRL) based cross-layer design in THz mesh backhaul networks (DEFLECT) is proposed, by considering dynamic traffic demands and possible sudden link failures. In DEFLECT, a heuristic routing metric is first devised to facilitate resource efficiency (RE) enhancement regarding energy and sub-array usages. Furthermore, a DRL based resource allocation algorithm is developed to realize long-term RE maximization and fast recovery from broken links. Specifically in the DRL method, the exploited multi-task structure cooperatively benefits joint power and sub-array allocation. Additionally, the leveraged hierarchical architecture realizes tailored resource allocation for each base station and learned knowledge transfer for fast recovery. Simulation results show that DEFLECT routing consumes less resource, compared to the minimal hop-count metric. Moreover, unlike conventional DRL methods causing packet loss and second-level latency, DEFLECT DRL realizes the long-term RE maximization with no packet loss and millisecond-level latency, and recovers resource-efficient backhaul from broken links within 1s.
Deep learning has shown remarkable success in the field of clustering recently. However, how to transfer a trained clustering model on a source domain to a target domain by leveraging the acquired knowledge to guide the clustering process remains challenging. Existing deep clustering methods often lack generalizability to new domains because they typically learn a group of fixed cluster centroids, which may not be optimal for the new domain distributions. In this paper, we propose a novel transferable deep clustering model that can automatically adapt the cluster centroids according to the distribution of data samples. Rather than learning a fixed set of centroids, our approach introduces a novel attention-based module that can adapt the centroids by measuring their relationship with samples. In addition, we theoretically show that our model is strictly more powerful than some classical clustering algorithms such as k-means or Gaussian Mixture Model (GMM). Experimental results on both synthetic and real-world datasets demonstrate the effectiveness and efficiency of our proposed transfer learning framework, which significantly improves the performance on target domain and reduces the computational cost.
A critical barrier to learning an accurate decision rule for outlier detection is the scarcity of outlier data. As such, practitioners often turn to the use of similar but imperfect outlier data from which they might transfer information to the target outlier detection task. Despite the recent empirical success of transfer learning approaches in outlier detection, a fundamental understanding of when and how knowledge can be transferred from a source to a target outlier detection task remains elusive. In this work, we adopt the traditional framework of Neyman-Pearson classification -- which formalizes supervised outlier detection -- with the added assumption that one has access to some related but imperfect outlier data. Our main results are as follows: We first determine the information-theoretic limits of the problem under a measure of discrepancy that extends some existing notions from traditional balanced classification; interestingly, unlike in balanced classification, seemingly very dissimilar sources can provide much information about a target, thus resulting in fast transfer. We then show that, in principle, these information-theoretic limits are achievable by adaptive procedures, i.e., procedures with no a priori information on the discrepancy between source and target outlier distributions.
This paper addresses challenges in robust transfer learning stemming from ambiguity in Bayes classifiers and weak transferable signals between the target and source distribution. We introduce a novel quantity called the ''ambiguity level'' that measures the discrepancy between the target and source regression functions, propose a simple transfer learning procedure, and establish a general theorem that shows how this new quantity is related to the transferability of learning in terms of risk improvements. Our proposed ''Transfer Around Boundary'' (TAB) model, with a threshold balancing the performance of target and source data, is shown to be both efficient and robust, improving classification while avoiding negative transfer. Moreover, we demonstrate the effectiveness of the TAB model on non-parametric classification and logistic regression tasks, achieving upper bounds which are optimal up to logarithmic factors. Simulation studies lend further support to the effectiveness of TAB. We also provide simple approaches to bound the excess misclassification error without the need for specialized knowledge in transfer learning.
The goal of our research is to automatically retrieve the satisfaction and the frustration in real-life call-center conversations. This study focuses an industrial application in which the customer satisfaction is continuously tracked down to improve customer services. To compensate the lack of large annotated emotional databases, we explore the use of pre-trained speech representations as a form of transfer learning towards AlloSat corpus. Moreover, several studies have pointed out that emotion can be detected not only in speech but also in facial trait, in biological response or in textual information. In the context of telephone conversations, we can break down the audio information into acoustic and linguistic by using the speech signal and its transcription. Our experiments confirms the large gain in performance obtained with the use of pre-trained features. Surprisingly, we found that the linguistic content is clearly the major contributor for the prediction of satisfaction and best generalizes to unseen data. Our experiments conclude to the definitive advantage of using CamemBERT representations, however the benefit of the fusion of acoustic and linguistic modalities is not as obvious. With models learnt on individual annotations, we found that fusion approaches are more robust to the subjectivity of the annotation task. This study also tackles the problem of performances variability and intends to estimate this variability from different views: weights initialization, confidence intervals and annotation subjectivity. A deep analysis on the linguistic content investigates interpretable factors able to explain the high contribution of the linguistic modality for this task.
Fine-tuning pre-trained Vision Transformers (ViTs) has showcased significant promise in enhancing visual recognition tasks. Yet, the demand for individualized and comprehensive fine-tuning processes for each task entails substantial computational and memory costs, posing a considerable challenge. Recent advancements in Parameter-Efficient Transfer Learning (PETL) have shown potential for achieving high performance with fewer parameter updates compared to full fine-tuning. However, their effectiveness is primarily observed in simple tasks like image classification, while they encounter challenges with more complex vision tasks like dense prediction. To address this gap, this study aims to identify an effective tuning method that caters to a wider range of visual tasks. In this paper, we introduce Hierarchical Side-Tuning (HST), an innovative PETL method facilitating the transfer of ViT models to diverse downstream tasks. Diverging from existing methods that focus solely on fine-tuning parameters within specific input spaces or modules, HST employs a lightweight Hierarchical Side Network (HSN). This network leverages intermediate activations from the ViT backbone to model multi-scale features, enhancing prediction capabilities. To evaluate HST, we conducted comprehensive experiments across a range of visual tasks, including classification, object detection, instance segmentation, and semantic segmentation. Remarkably, HST achieved state-of-the-art performance in 13 out of the 19 tasks on the VTAB-1K benchmark, with the highest average Top-1 accuracy of 76.1%, while fine-tuning a mere 0.78M parameters. When applied to object detection and semantic segmentation tasks on the COCO and ADE20K testdev benchmarks, HST outperformed existing PETL methods and even surpassed full fine-tuning.