Jaekyun Moon

Dong-Jun Han, Do-Yeon Kim, Minseok Choi et al.

A fundamental challenge to providing edge-AI services is the need for a machine learning (ML) model that achieves personalization (i.e., to individual clients) and generalization (i.e., to unseen data) properties concurrently. Existing techniques in federated learning (FL) have encountered a steep tradeoff between these objectives and impose large computational requirements on edge devices during training and inference. In this paper, we propose SplitGP, a new split learning solution that can simultaneously capture generalization and personalization capabilities for efficient inference across resource-constrained clients (e.g., mobile/IoT devices). Our key idea is to split the full ML model into client-side and server-side components, and impose different roles to them: the client-side model is trained to have strong personalization capability optimized to each client's main task, while the server-side model is trained to have strong generalization capability for handling all clients' out-of-distribution tasks. We analytically characterize the convergence behavior of SplitGP, revealing that all client models approach stationary points asymptotically. Further, we analyze the inference time in SplitGP and provide bounds for determining model split ratios. Experimental results show that SplitGP outperforms existing baselines by wide margins in inference time and test accuracy for varying amounts of out-of-distribution samples.

Jungwuk Park, Dong-Jun Han, Soyeong Kim et al.

In domain generalization (DG), the target domain is unknown when the model is being trained, and the trained model should successfully work on an arbitrary (and possibly unseen) target domain during inference. This is a difficult problem, and despite active studies in recent years, it remains a great challenge. In this paper, we take a simple yet effective approach to tackle this issue. We propose test-time style shifting, which shifts the style of the test sample (that has a large style gap with the source domains) to the nearest source domain that the model is already familiar with, before making the prediction. This strategy enables the model to handle any target domains with arbitrary style statistics, without additional model update at test-time. Additionally, we propose style balancing, which provides a great platform for maximizing the advantage of test-time style shifting by handling the DG-specific imbalance issues. The proposed ideas are easy to implement and successfully work in conjunction with various other DG schemes. Experimental results on different datasets show the effectiveness of our methods.

Daejung Yoon, Jaekyun Moon

We consider channel estimation specific to turbo equalization for multiple-input multiple-output (MIMO) wireless communication. We develop a soft-decision-driven sequential algorithm geared to the pipelined turbo equalizer architecture operating on orthogonal frequency division multiplexing (OFDM) symbols. One interesting feature of the pipelined turbo equalizer is that multiple soft-decisions become available at various processing stages. A tricky issue is that these multiple decisions from different pipeline stages have varying levels of reliability. This paper establishes an effective strategy for the channel estimator to track the target channel, while dealing with observation sets with different qualities. The resulting algorithm is basically a linear sequential estimation algorithm and, as such, is Kalman-based in nature. The main difference here, however, is that the proposed algorithm employs puncturing on observation samples to effectively deal with the inherent correlation among the multiple demapper/decoder module outputs that cannot easily be removed by the traditional innovations approach. The proposed algorithm continuously monitors the quality of the feedback decisions and incorporates it in the channel estimation process. The proposed channel estimation scheme shows clear performance advantages relative to existing channel estimation techniques.

Jungwuk Park, Dong-Jun Han, Jinho Kim et al.

Traditional federated learning (FL) algorithms operate under the assumption that the data distributions at training (source domains) and testing (target domain) are the same. The fact that domain shifts often occur in practice necessitates equipping FL methods with a domain generalization (DG) capability. However, existing DG algorithms face fundamental challenges in FL setups due to the lack of samples/domains in each client's local dataset. In this paper, we propose StableFDG, a style and attention based learning strategy for accomplishing federated domain generalization, introducing two key contributions. The first is style-based learning, which enables each client to explore novel styles beyond the original source domains in its local dataset, improving domain diversity based on the proposed style sharing, shifting, and exploration strategies. Our second contribution is an attention-based feature highlighter, which captures the similarities between the features of data samples in the same class, and emphasizes the important/common characteristics to better learn the domain-invariant characteristics of each class in data-poor FL scenarios. Experimental results show that StableFDG outperforms existing baselines on various DG benchmark datasets, demonstrating its efficacy.

Seokil Ham, Jungwuk Park, Dong-Jun Han et al.

While multi-exit neural networks are regarded as a promising solution for making efficient inference via early exits, combating adversarial attacks remains a challenging problem. In multi-exit networks, due to the high dependency among different submodels, an adversarial example targeting a specific exit not only degrades the performance of the target exit but also reduces the performance of all other exits concurrently. This makes multi-exit networks highly vulnerable to simple adversarial attacks. In this paper, we propose NEO-KD, a knowledge-distillation-based adversarial training strategy that tackles this fundamental challenge based on two key contributions. NEO-KD first resorts to neighbor knowledge distillation to guide the output of the adversarial examples to tend to the ensemble outputs of neighbor exits of clean data. NEO-KD also employs exit-wise orthogonal knowledge distillation for reducing adversarial transferability across different submodels. The result is a significantly improved robustness against adversarial attacks. Experimental results on various datasets/models show that our method achieves the best adversarial accuracy with reduced computation budgets, compared to the baselines relying on existing adversarial training or knowledge distillation techniques for multi-exit networks.

Hasnain Irshad Bhatti, Jaekyun Moon

Locally supervised learning aims to train a neural network based on a local estimation of the global loss function at each decoupled module of the network. Auxiliary networks are typically appended to the modules to approximate the gradient updates based on the greedy local losses. Despite being advantageous in terms of parallelism and reduced memory consumption, this paradigm of training severely degrades the generalization performance of neural networks. In this paper, we propose Periodically Guided local Learning (PGL), which reinstates the global objective repetitively into the local-loss based training of neural networks primarily to enhance the model's generalization capability. We show that a simple periodic guidance scheme begets significant performance gains while having a low memory footprint. We conduct extensive experiments on various datasets and networks to demonstrate the effectiveness of PGL, especially in the configuration with numerous decoupled modules.

Mohammad Mahdi Rahimi, Hasnain Irshad Bhatti, Younghyun Park et al.

Federated Learning (FL) is a decentralized machine learning paradigm that enables collaborative model training across dispersed nodes without having to force individual nodes to share data. However, its broad adoption is hindered by the high communication costs of transmitting a large number of model parameters. This paper presents EvoFed, a novel approach that integrates Evolutionary Strategies (ES) with FL to address these challenges. EvoFed employs a concept of 'fitness-based information sharing', deviating significantly from the conventional model-based FL. Rather than exchanging the actual updated model parameters, each node transmits a distance-based similarity measure between the locally updated model and each member of the noise-perturbed model population. Each node, as well as the server, generates an identical population set of perturbed models in a completely synchronized fashion using the same random seeds. With properly chosen noise variance and population size, perturbed models can be combined to closely reflect the actual model updated using the local dataset, allowing the transmitted similarity measures (or fitness values) to carry nearly the complete information about the model parameters. As the population size is typically much smaller than the number of model parameters, the savings in communication load is large. The server aggregates these fitness values and is able to update the global model. This global fitness vector is then disseminated back to the nodes, each of which applies the same update to be synchronized to the global model. Our analysis shows that EvoFed converges, and our experimental results validate that at the cost of increased local processing loads, EvoFed achieves performance comparable to FedAvg while reducing overall communication requirements drastically in various practical settings.

Humaira Kousar, Hasnain Irshad Bhatti, Jaekyun Moon

Efficient data selection is crucial for enhancing the training efficiency of deep neural networks and minimizing annotation requirements. Traditional methods often face high computational costs, limiting their scalability and practical use. We introduce PruneFuse, a novel strategy that leverages pruned networks for data selection and later fuses them with the original network to optimize training. PruneFuse operates in two stages: First, it applies structured pruning to create a smaller pruned network that, due to its structural coherence with the original network, is well-suited for the data selection task. This small network is then trained and selects the most informative samples from the dataset. Second, the trained pruned network is seamlessly fused with the original network. This integration leverages the insights gained during the training of the pruned network to facilitate the learning process of the fused network while leaving room for the network to discover more robust solutions. Extensive experimentation on various datasets demonstrates that PruneFuse significantly reduces computational costs for data selection, achieves better performance than baselines, and accelerates the overall training process.

Wonjeong Choi, Jungwuk Park, Dong-Jun Han et al.

Research interests in the robustness of deep neural networks against domain shifts have been rapidly increasing in recent years. Most existing works, however, focus on improving the accuracy of the model, not the calibration performance which is another important requirement for trustworthy AI systems. Temperature scaling (TS), an accuracy-preserving post-hoc calibration method, has been proven to be effective in in-domain settings, but not in out-of-domain (OOD) due to the difficulty in obtaining a validation set for the unseen domain beforehand. In this paper, we propose consistency-guided temperature scaling (CTS), a new temperature scaling strategy that can significantly enhance the OOD calibration performance by providing mutual supervision among data samples in the source domains. Motivated by our observation that over-confidence stemming from inconsistent sample predictions is the main obstacle to OOD calibration, we propose to guide the scaling process by taking consistencies into account in terms of two different aspects -- style and content -- which are the key components that can well-represent data samples in multi-domain settings. Experimental results demonstrate that our proposed strategy outperforms existing works, achieving superior OOD calibration performance on various datasets. This can be accomplished by employing only the source domains without compromising accuracy, making our scheme directly applicable to various trustworthy AI systems.

Humaira Kousar, Hasnain Irshad Bhatti, Jaekyun Moon

Efficient data selection is essential for improving the training efficiency of deep neural networks and reducing the associated annotation costs. However, traditional methods tend to be computationally expensive, limiting their scalability and real-world applicability. We introduce PruneFuse, a novel method that combines pruning and network fusion to enhance data selection and accelerate network training. In PruneFuse, the original dense network is pruned to generate a smaller surrogate model that efficiently selects the most informative samples from the dataset. Once this iterative data selection selects sufficient samples, the insights learned from the pruned model are seamlessly integrated with the dense model through network fusion, providing an optimized initialization that accelerates training. Extensive experimentation on various datasets demonstrates that PruneFuse significantly reduces computational costs for data selection, achieves better performance than baselines, and accelerates the overall training process.

Jun Seo, Young-Hyun Park, Sung Whan Yoon et al.

Few-shot learning allows machines to classify novel classes using only a few labeled samples. Recently, few-shot segmentation aiming at semantic segmentation on low sample data has also seen great interest. In this paper, we propose a learnable module that can be placed on top of existing segmentation networks for performing few-shot segmentation. This module, called the task-adaptive feature transformer (TAFT), linearly transforms task-specific high-level features to a set of task agnostic features well-suited to conducting few-shot segmentation. The task-conditioned feature transformation allows an effective utilization of the semantic information in novel classes to generate tight segmentation masks. We also propose a semantic enrichment (SE) module that utilizes a pixel-wise attention module for high-level feature and an auxiliary loss from an auxiliary segmentation network conducting the semantic segmentation for all training classes. Experiments on PASCAL-$5^i$ and COCO-$20^i$ datasets confirm that the added modules successfully extend the capability of existing segmentators to yield highly competitive few-shot segmentation performances.

Jy-yong Sohn, Liang Shang, Hongxu Chen et al.

Mixup is a data augmentation method that generates new data points by mixing a pair of input data. While mixup generally improves the prediction performance, it sometimes degrades the performance. In this paper, we first identify the main causes of this phenomenon by theoretically and empirically analyzing the mixup algorithm. To resolve this, we propose GenLabel, a simple yet effective relabeling algorithm designed for mixup. In particular, GenLabel helps the mixup algorithm correctly label mixup samples by learning the class-conditional data distribution using generative models. Via extensive theoretical and empirical analysis, we show that mixup, when used together with GenLabel, can effectively resolve the aforementioned phenomenon, improving the generalization performance and the adversarial robustness.

Beongjun Choi, Jy-yong Sohn, Dong-Jun Han et al.

Federated learning has been spotlighted as a way to train neural networks using distributed data with no need for individual nodes to share data. Unfortunately, it has also been shown that adversaries may be able to extract local data contents off model parameters transmitted during federated learning. A recent solution based on the secure aggregation primitive enabled privacy-preserving federated learning, but at the expense of significant extra communication/computational resources. In this paper, we propose a low-complexity scheme that provides data privacy using substantially reduced communication/computational resources relative to the existing secure solution. The key idea behind the suggested scheme is to design the topology of secret-sharing nodes as a sparse random graph instead of the complete graph corresponding to the existing solution. We first obtain the necessary and sufficient condition on the graph to guarantee both reliability and privacy. We then suggest using the Erdős-Rényi graph in particular and provide theoretical guarantees on the reliability/privacy of the proposed scheme. Through extensive real-world experiments, we demonstrate that our scheme, using only $20 \sim 30\%$ of the resources required in the conventional scheme, maintains virtually the same levels of reliability and data privacy in practical federated learning systems.

Jun Seo, Young-Hyun Park, Sung-Whan Yoon et al.

Few-shot learning allows machines to classify novel classes using only a few labeled samples. Recently, few-shot segmentation aiming at semantic segmentation on low sample data has also seen great interest. In this paper, we propose a learnable module for few-shot segmentation, the task-adaptive feature transformer (TAFT). TAFT linearly transforms task-specific high-level features to a set of task-agnostic features well-suited to the segmentation job. Using this task-conditioned feature transformation, the model is shown to effectively utilize the semantic information in novel classes to generate tight segmentation masks. The proposed TAFT module can be easily plugged into existing semantic segmentation algorithms to achieve few-shot segmentation capability with only a few added parameters. We combine TAFT with Deeplab V3+, a well-known segmentation architecture; experiments on the PASCAL-$5^i$ dataset confirm that this combination successfully adds few-shot learning capability to the segmentation algorithm, achieving the state-of-the-art few-shot segmentation performance in some key representative cases.

Sung Whan Yoon, Do-Yeon Kim, Jun Seo et al.

Learning novel concepts while preserving prior knowledge is a long-standing challenge in machine learning. The challenge gets greater when a novel task is given with only a few labeled examples, a problem known as incremental few-shot learning. We propose XtarNet, which learns to extract task-adaptive representation (TAR) for facilitating incremental few-shot learning. The method utilizes a backbone network pretrained on a set of base categories while also employing additional modules that are meta-trained across episodes. Given a new task, the novel feature extracted from the meta-trained modules is mixed with the base feature obtained from the pretrained model. The process of combining two different features provides TAR and is also controlled by meta-trained modules. The TAR contains effective information for classifying both novel and base categories. The base and novel classifiers quickly adapt to a given task by utilizing the TAR. Experiments on standard image datasets indicate that XtarNet achieves state-of-the-art incremental few-shot learning performance. The concept of TAR can also be used in conjunction with existing incremental few-shot learning methods; extensive simulation results in fact show that applying TAR enhances the known methods significantly.

Young-Hyun Park, Jun Seo, Jaekyun Moon

We tackle a novel few-shot learning challenge, which we call few-shot semantic edge detection, aiming to localize crisp boundaries of novel categories using only a few labeled samples. We also present a Class-Agnostic Few-shot Edge detection Network (CAFENet) based on meta-learning strategy. CAFENet employs a semantic segmentation module in small-scale to compensate for lack of semantic information in edge labels. The predicted segmentation mask is used to generate an attention map to highlight the target object region, and make the decoder module concentrate on that region. We also propose a new regularization method based on multi-split matching. In meta-training, the metric-learning problem with high-dimensional vectors are divided into small subproblems with low-dimensional sub-vectors. Since there is no existing dataset for few-shot semantic edge detection, we construct two new datasets, FSE-1000 and SBD-$5^i$, and evaluate the performance of the proposed CAFENet on them. Extensive simulation results confirm the performance merits of the techniques adopted in CAFENet.

Jun Seo, Sung Whan Yoon, Jaekyun Moon

Few-shot learning aims to handle previously unseen tasks using only a small amount of new training data. In preparing (or meta-training) a few-shot learner, however, massive labeled data are necessary. In the real world, unfortunately, labeled data are expensive and/or scarce. In this work, we propose a few-shot learner that can work well under the semi-supervised setting where a large portion of training data is unlabeled. Our method employs explicit task-conditioning in which unlabeled sample clustering for the current task takes place in a new projection space different from the embedding feature space. The conditioned clustering space is linearly constructed so as to quickly close the gap between the class centroids for the current task and the independent per-class reference vectors meta-trained across tasks. In a more general setting, our method introduces a concept of controlling the degree of task-conditioning for meta-learning: the amount of task-conditioning varies with the number of repetitive updates for the clustering space. Extensive simulation results based on the miniImageNet and tieredImageNet datasets show state-of-the-art semi-supervised few-shot classification performance of the proposed method. Simulation results also indicate that the proposed task-adaptive clustering shows graceful degradation with a growing number of distractor samples, i.e., unlabeled sample images coming from outside the candidate classes.

Jy-yong Sohn, Dong-Jun Han, Beongjun Choi et al.

Recent advances in large-scale distributed learning algorithms have enabled communication-efficient training via SignSGD. Unfortunately, a major issue continues to plague distributed learning: namely, Byzantine failures may incur serious degradation in learning accuracy. This paper proposes Election Coding, a coding-theoretic framework to guarantee Byzantine-robustness for SignSGD with Majority Vote, which uses minimum worker-master communication in both directions. The suggested framework explores new information-theoretic limits of finding the majority opinion when some workers could be malicious, and paves the road to implement robust and efficient distributed learning algorithms. Under this framework, we construct two types of explicit codes, random Bernoulli codes and deterministic algebraic codes, that can tolerate Byzantine attacks with a controlled amount of computational redundancy. For the Bernoulli codes, we provide upper bounds on the error probability in estimating the majority opinion, which give useful insights into code design for tolerating Byzantine attacks. As for deterministic codes, we construct an explicit code which perfectly tolerates Byzantines, and provide tight upper/lower bounds on the minimum required computational redundancy. Finally, the Byzantine-tolerance of the suggested coding schemes is confirmed by deep learning experiments on Amazon EC2 using Python with MPI4py package.

Sung Whan Yoon, Jun Seo, Jaekyun Moon

Handling previously unseen tasks after given only a few training examples continues to be a tough challenge in machine learning. We propose TapNets, neural networks augmented with task-adaptive projection for improved few-shot learning. Here, employing a meta-learning strategy with episode-based training, a network and a set of per-class reference vectors are learned across widely varying tasks. At the same time, for every episode, features in the embedding space are linearly projected into a new space as a form of quick task-specific conditioning. The training loss is obtained based on a distance metric between the query and the reference vectors in the projection space. Excellent generalization results in this way. When tested on the Omniglot, miniImageNet and tieredImageNet datasets, we obtain state of the art classification accuracies under various few-shot scenarios.

Sung Whan Yoon, Jun Seo, Jaekyun Moon

We propose a meta-learning algorithm utilizing a linear transformer that carries out null-space projection of neural network outputs. The main idea is to construct an alternative classification space such that the error signals during few-shot learning are quickly zero-forced on that space so that reliable classification on low data is possible. The final decision on a query is obtained utilizing a null-space-projected distance measure between the network output and reference vectors, both of which have been trained in the initial learning phase. Among the known methods with a given model size, our meta-learner achieves the best or near-best image classification accuracies with Omniglot and miniImageNet datasets.