Chuan-Sheng Foo

ICLR Conference 2025 Conference Paper

Broaden your SCOPE! Efficient Multi-turn Conversation Planning for LLMs with Semantic Space

• Zhiliang Chen
• Xinyuan Niu 0001
• Chuan-Sheng Foo
• Bryan Kian Hsiang Low

Large language models (LLMs) are used in chatbots or AI assistants to hold conversations with a human user. In such applications, the quality (e.g., user engagement, safety) of a conversation is important and can only be exactly known at the end of the conversation. To maximize its expected quality, conversation planning reasons about the stochastic transitions within a conversation to select the optimal LLM response at each turn. Existing simulation-based conversation planning algorithms typically select the optimal response by simulating future conversations with a large number of LLM queries at every turn. However, this process is extremely time-consuming and hence impractical for real-time conversations. This paper presents a novel approach called Semantic space COnversation Planning with improved Efficiency (SCOPE) that exploits the dense semantic representation of conversations to perform conversation planning efficiently. In particular, SCOPE models the stochastic transitions in conversation semantics and their associated rewards to plan entirely within the semantic space. This allows us to select the optimal LLM response at every conversation turn without needing additional LLM queries for simulation. As a result, SCOPE can perform conversation planning 70 times faster than conventional simulation-based planning algorithms when applied to a wide variety of conversation starters and two reward functions seen in the real world, yet achieving a higher reward within a practical planning budget. Our code can be found at: https://github.com/chenzhiliang94/convo-plan-SCOPE.

ICLR Conference 2025 Conference Paper

Confidence Elicitation: A New Attack Vector for Large Language Models

• Brian Formento
• Chuan-Sheng Foo
• See-Kiong Ng

A fundamental issue in deep learning has been adversarial robustness. As these systems have scaled, such issues have persisted. Currently, large language models (LLMs) with billions of parameters suffer from adversarial attacks just like their earlier, smaller counterparts. However, the threat models have changed. Previously, having gray-box access, where input embeddings or output logits/probabilities were visible to the user, might have been reasonable. However, with the introduction of closed-source models, no information about the model is available apart from the generated output. This means that current black-box attacks can only utilize the final prediction to detect if an attack is successful. In this work, we investigate and demonstrate the potential of attack guidance, akin to using output probabilities, while having only black-box access in a classification setting. This is achieved through the ability to elicit confidence from the model. We empirically show that the elicited confidence is calibrated and not hallucinated for current LLMs. By minimizing the elicited confidence, we can therefore increase the likelihood of misclassification. Our new proposed paradigm demonstrates promising state-of-the-art results on three datasets across two models (LLaMA-3-8B-Instruct and Mistral-7B-Instruct-V0.3) when comparing our technique to existing hard-label black-box attack methods that introduce word-level substitutions. The code is publicly available at GitHub: Confidence_Elicitation_Attacks.

ICLR Conference 2025 Conference Paper

Evidential Learning-based Certainty Estimation for Robust Dense Feature Matching

• Lile Cai
• Chuan-Sheng Foo
• Xun Xu 0002
• Zaiwang Gu
• Jun Cheng 0003
• Xulei Yang

Dense feature matching methods aim to estimate a dense correspondence field between images. Inaccurate correspondence can occur due to the presence of unmatchable region, necessitating the need for certainty measurement. This is typically addressed by training a binary classifier to decide whether each predicted correspondence is reliable. However, deep neural network-based classifiers can be vulnerable to image corruptions or perturbations, making it difficult to obtain reliable matching pairs in corrupted scenario. In this work, we propose an evidential deep learning framework to enhance the robustness of dense matching against corruptions. We modify the certainty prediction branch in dense matching models to generate appropriate belief masses and compute the certainty score by taking expectation over the resulting Dirichlet distribution. We evaluate our method on a wide range of benchmarks and show that our method leads to improved robustness against common corruptions and adversarial attacks, achieving up to 10.1\% improvement under severe corruptions.

ICML Conference 2025 Conference Paper

NICE Data Selection for Instruction Tuning in LLMs with Non-differentiable Evaluation Metric

• Jingtan Wang 0001
• Xiaoqiang Lin
• Rui Qiao 0006
• Pang Wei Koh
• Chuan-Sheng Foo
• Bryan Kian Hsiang Low

Curating data for instruction tuning is crucial for enhancing the performance of large language models (LLMs). This work aims to select training data for instruction tuning to improve the LLM performance on specific tasks. Existing methods often rely on next-token prediction (NTP) loss as a proxy for target task performance due to the non-differentiable nature of performance evaluation metrics. They select training data points that are most helpful in reducing validation loss. However, there is a discrepancy between minimizing NTP loss and maximizing performance (e. g. , code pass rate in code generation). To remedy this, we introduce a novel Non-differentiable evaluation metric-based InfluenCe Estimation (NICE), which leverages the policy gradient to select the training data that improves the performance. Moreover, NICE can perform data selection in the absence of labels (ground-truth responses) when the evaluation metrics do not require labels (e. g. , a reward model can output reward scores without supervision from labels). Experimental results show that our approach outperforms existing data selection baselines that use NTP loss in diverse and realistic scenarios. Notably, subsets selected by NICE often produce models that outperform those trained on the full dataset. Our code is available at https: //github. com/JTWang2000/NICE.

ICLR Conference 2025 Conference Paper

On the Adversarial Risk of Test Time Adaptation: An Investigation into Realistic Test-Time Data Poisoning

• Yongyi Su
• Yushu Li
• Nanqing Liu
• Kui Jia
• Xulei Yang
• Chuan-Sheng Foo
• Xun Xu 0002

Test-time adaptation (TTA) updates the model weights during the inference stage using testing data to enhance generalization. However, this practice exposes TTA to adversarial risks. Existing studies have shown that when TTA is updated with crafted adversarial test samples, also known as test-time poisoned data, the performance on benign samples can deteriorate. Nonetheless, the perceived adversarial risk may be overstated if the poisoned data is generated under overly strong assumptions. In this work, we first review realistic assumptions for test-time data poisoning, including white-box versus grey-box attacks, access to benign data, attack order, and more. We then propose an effective and realistic attack method that better produces poisoned samples without access to benign samples, and derive an effective in-distribution attack objective. We also design two TTA-aware attack objectives. Our benchmarks of existing attack methods reveal that the TTA methods are more robust than previously believed. In addition, we analyze effective defense strategies to help develop adversarially robust TTA methods. The source code is available at https://github.com/Gorilla-Lab-SCUT/RTTDP.

ICLR Conference 2025 Conference Paper

Robust-PIFu: Robust Pixel-aligned Implicit Function for 3D Human Digitalization from a Single Image

• Kennard Yanting Chan
• Fayao Liu
• Guosheng Lin
• Chuan-Sheng Foo
• Weisi Lin

Existing methods for 3D clothed human digitalization perform well when the input image is captured in ideal conditions that assume the lack of any occlusion. However, in reality, images may often have occlusion problems such as incomplete observation of the human subject's full body, self-occlusion by the human subject, and non-frontal body pose. When given such input images, these existing methods fail to perform adequately. Thus, we propose Robust-PIFu, a pixel-aligned implicit model that capitalized on large-scale, pretrained latent diffusion models to address the challenge of digitalizing human subjects from non-ideal images that suffer from occlusions. Robust-PIfu offers four new contributions. Firstly, we propose a 'disentangling' latent diffusion model. This diffusion model, pretrained on billions of images, takes in any input image and removes external occlusions, such as inter-person occlusions, from that image. Secondly, Robust-PIFu addresses internal occlusions like self-occlusion by introducing a `penetrating' latent diffusion model. This diffusion model outputs multi-layered normal maps that by-pass occlusions caused by the human subject's own limbs or other body parts (i.e. self-occlusion). Thirdly, in order to incorporate such multi-layered normal maps into a pixel-aligned implicit model, we introduce our Layered-Normals Pixel-aligned Implicit Model, which improves the structural accuracy of predicted clothed human meshes. Lastly, Robust-PIFu proposes an optional super-resolution mechanism for the multi-layered normal maps. This addresses scenarios where the input image is of low or inadequate resolution. Though not strictly related to occlusion, this is still an important subproblem. Our experiments show that Robust-PIFu outperforms current SOTA methods both qualitatively and quantitatively. Our code will be released to the public.

ICLR Conference 2024 Conference Paper

A Unified Framework for Bayesian Optimization under Contextual Uncertainty

• Sebastian Shenghong Tay
• Chuan-Sheng Foo
• Daisuke Urano
• Richalynn Leong
• Bryan Kian Hsiang Low

Bayesian optimization under contextual uncertainty (BOCU) is a family of BO problems in which the learner makes a decision prior to observing the context and must manage the risks involved. Distributionally robust BO (DRBO) is a subset of BOCU that affords robustness against context distribution shift, and includes the optimization of expected values and worst-case values as special cases. By considering the first derivatives of the DRBO objective, we generalize DRBO to one that includes several other uncertainty objectives studied in the BOCU literature such as worst-case sensitivity (and thus notions of risk such as variance, range, and conditional value-at-risk) and mean-risk tradeoffs. We develop a general Thompson sampling algorithm that is able to optimize any objective within the BOCU framework, analyze its theoretical properties, and compare it to suitable baselines across different experimental settings and uncertainty objectives.

NeurIPS Conference 2024 Conference Paper

Data Distribution Valuation

• Xinyi Xu
• Shuaiqi Wang
• Chuan-Sheng Foo
• Bryan K. Low
• Giulia Fanti

Data valuation is a class of techniques for quantitatively assessing the value of data for applications like pricing in data marketplaces. Existing data valuation methods define a value for a discrete dataset. However, in many use cases, users are interested in not only the value of the dataset, but that of the distribution from which the dataset was sampled. For example, consider a buyer trying to evaluate whether to purchase data from different vendors. The buyer may observe (and compare) only a small preview sample from each vendor, to decide which vendor's data distribution is most useful to the buyer and purchase. The core question is how should we compare the values of data distributions from their samples? Under a Huber characterization of the data heterogeneity across vendors, we propose a maximum mean discrepancy (MMD)-based valuation method which enables theoretically principled and actionable policies for comparing data distributions from samples. We empirically demonstrate that our method is sample-efficient and effective in identifying valuable data distributions against several existing baselines, on multiple real-world datasets (e. g. , network intrusion detection, credit card fraud detection) and downstream applications (classification, regression).

ICML Conference 2024 Conference Paper

Helpful or Harmful Data? Fine-tuning-free Shapley Attribution for Explaining Language Model Predictions

• Jingtan Wang 0001
• Xiaoqiang Lin
• Rui Qiao 0006
• Chuan-Sheng Foo
• Bryan Kian Hsiang Low

The increasing complexity of foundational models underscores the necessity for explainability, particularly for fine-tuning, the most widely used training method for adapting models to downstream tasks. Instance attribution, one type of explanation, attributes the model prediction to each training example by an instance score. However, the robustness of instance scores, specifically towards dataset resampling, has been overlooked. To bridge this gap, we propose a notion of robustness on the sign of the instance score. We theoretically and empirically demonstrate that the popular leave-one-out-based methods lack robustness, while the Shapley value behaves significantly better, but at a higher computational cost. Accordingly, we introduce an efficient fine-tuning-free approximation of the Shapley value (FreeShap) for instance attribution based on the neural tangent kernel. We empirically demonstrate that FreeShap outperforms other methods for instance attribution and other data-centric applications such as data removal, data selection, and wrong label detection, and further generalize our scale to large language models (LLMs). Our code is available at https: //github. com/JTWang2000/FreeShap.

TMLR Journal 2024 Journal Article

Hybrid Active Learning with Uncertainty-Weighted Embeddings

• Yinan He
• Lile Cai
• Jingyi Liao
• Chuan-Sheng Foo

We introduce a hybrid active learning method that simultaneously considers uncertainty and diversity for sample selection. Our method consists of two key steps: computing a novel uncertainty-weighted embedding, then applying distance-based sampling for sample selection. Our proposed uncertainty-weighted embedding is computed by weighting a sample's feature representation by an uncertainty measure. We show how this embedding generalizes the gradient embedding of BADGE so it can be used with arbitrary loss functions and be computed more efficiently, especially for dense prediction tasks and network architectures with large numbers of parameters in the final layer. We extensively evaluate the proposed hybrid active learning method on image classification, semantic segmentation and object detection tasks, and demonstrate that it achieves state-of-the-art performance.

ICML Conference 2024 Conference Paper

Pseudo-Calibration: Improving Predictive Uncertainty Estimation in Unsupervised Domain Adaptation

• Dapeng Hu
• Jian Liang 0001
• Xinchao Wang
• Chuan-Sheng Foo

Unsupervised domain adaptation (UDA) has seen substantial efforts to improve model accuracy for an unlabeled target domain with the help of a labeled source domain. However, UDA models often exhibit poorly calibrated predictive uncertainty on target data, a problem that remains under-explored and poses risks in safety-critical UDA applications. The calibration problem in UDA is particularly challenging due to the absence of labeled target data and severe distribution shifts between domains. In this paper, we approach UDA calibration as a target-domain-specific unsupervised problem, different from mainstream solutions based on covariate shift. We introduce Pseudo-Calibration (PseudoCal), a novel post-hoc calibration framework. Our innovative use of inference-stage mixup synthesizes a labeled pseudo-target set capturing the structure of the real unlabeled target data. This turns the unsupervised calibration problem into a supervised one, easily solvable with temperature scaling. Extensive empirical evaluations across 5 diverse UDA scenarios involving 10 UDA methods consistently demonstrate the superior performance and versatility of PseudoCal over existing solutions.

ICML Conference 2024 Conference Paper

Towards AutoAI: Optimizing a Machine Learning System with Black-box and Differentiable Components

• Zhiliang Chen
• Chuan-Sheng Foo
• Bryan Kian Hsiang Low

Machine learning (ML) models in the real world typically do not exist in isolation. They are usually part of a complex system (e. g. , healthcare systems, self-driving cars) containing multiple ML and black-box components. The problem of optimizing such systems, which we refer to as automated AI (AutoAI), requires us to jointly train all ML components together and presents a significant challenge because the number of system parameters is extremely high and the system has no analytical form. To circumvent this, we introduce a novel algorithm called A-BAD-BO which uses each ML component’s local loss as an auxiliary indicator for system performance. A-BAD-BO uses Bayesian optimization (BO) to optimize the local loss configuration of a system in a smaller dimensional space and exploits the differentiable structure of ML components to recover optimal system parameters from the optimized configuration. We show A-BAD-BO converges to optimal system parameters by showing that it is asymptotically no regret. We use A-BAD-BO to optimize several synthetic and real-world complex systems, including a prompt engineering pipeline for large language models containing millions of system parameters. Our results demonstrate that A-BAD-BO yields better system optimality than gradient-driven baselines and is more sample-efficient than pure BO algorithms.

NeurIPS Conference 2024 Conference Paper

Towards Reliable Model Selection for Unsupervised Domain Adaptation: An Empirical Study and A Certified Baseline

• Dapeng Hu
• Mi Luo
• Jian Liang
• Chuan-Sheng Foo

Selecting appropriate hyperparameters is crucial for unlocking the full potential of advanced unsupervised domain adaptation (UDA) methods in unlabeled target domains. Although this challenge remains under-explored, it has recently garnered increasing attention with the proposals of various model selection methods. Reliable model selection should maintain performance across diverse UDA methods and scenarios, especially avoiding highly risky worst-case selections—selecting the model or hyperparameter with the worst performance in the pool. \textit{Are existing model selection methods reliable and versatile enough for different UDA tasks? } In this paper, we provide a comprehensive empirical study involving 8 existing model selection approaches to answer this question. Our evaluation spans 12 UDA methods across 5 diverse UDA benchmarks and 5 popular UDA scenarios. Surprisingly, we find that none of these approaches can effectively avoid the worst-case selection. In contrast, a simple but overlooked ensemble-based selection approach, which we call EnsV, is both theoretically and empirically certified to avoid the worst-case selection, ensuring high reliability. Additionally, EnsV is versatile for various practical but challenging UDA scenarios, including validation of open-partial-set UDA and source-free UDA. Finally, we call for more attention to the reliability of model selection in UDA: avoiding the worst-case is as significant as achieving peak selection performance and should not be overlooked when developing new model selection methods. Code is available at https: //github. com/LHXXHB/EnsV.

TMLR Journal 2024 Journal Article

Uniformly Distributed Feature Representations for Fair and Robust Learning

• Kiran Krishnamachari
• See-Kiong Ng
• Chuan-Sheng Foo

A fundamental challenge in machine learning is training models that generalize well to distributions different from the training distribution. Empirical Risk Minimization (ERM), which is the predominant learning principle, is known to under-perform in minority sub-populations and fail to generalize well in unseen test domains. In this work, we propose a novel learning principle called Uniform Risk Minimization (URM) to alleviate these issues. We first show theoretically that uniform training data distributions and feature representations support robustness to distribution shifts. Motivated by this result, we propose an empirical method that trains deep neural networks to learn a uniformly distributed feature representation in their final activation layer for improved robustness. Our experiments on multiple datasets for sub-population shifts and domain generalization show that URM improves the generalization of deep neural networks without requiring knowledge of groups or domains during training. URM is competitive with the best existing methods designed for these tasks and can also be easily combined with them for improved performance. Our work sheds light on the importance of the distribution of learned feature representations for model robustness and fairness. Code is available at https://github.com/kiranchari/UniformRiskMinimization.

TMLR Journal 2024 Journal Article

When low-vision task meets dense prediction tasks with less data: an auxiliary self-trained geometry regularization

• Zaiwang Gu
• Weide Liu
• Xulei Yang
• Chuan-Sheng Foo
• Jun Cheng

Many deep learning methods are data-driven, often converging to local minima due to limited training data. This situation poses a challenge in domains where acquiring adequate data is difficult for model training or fine-tuning, such as generalized few-shot semantic segmentation (GFSSeg) and monocular depth estimation (MDE). To this end, we propose a self-trained geometry regularization framework to enhance model training or fine-tuning in scenarios with limited training data using geometric knowledge. Specifically, we propose to leverage low-level geometry information extracted from the training data and define a novel regularization term, which is a plug-and-play module jointly trained with the primary task via multi-task learning. Our proposed regularization neither relies on extra manual labels and data in training nor requires extra computation during the inference stage. We demonstrate the effectiveness of this regularization on GFSSeg and MDE tasks. Notably, it improves the state-of-the-art GFSSeg by 5.61% and 4.26% mIoU of novel classes on PASCAL and COCO in the 1-shot scenario. In MDE, it achieves a relative reduction of SILog error by 16.6% and 9.4% for two recent methods in the KITTI dataset.

ICML Conference 2023 Conference Paper

Fair yet Asymptotically Equal Collaborative Learning

• Xiaoqiang Lin
• Xinyi Xu
• See-Kiong Ng
• Chuan-Sheng Foo
• Bryan Kian Hsiang Low

In collaborative learning with streaming data, nodes (e. g. , organizations) jointly and continuously learn a machine learning (ML) model by sharing the latest model updates computed from their latest streaming data. For the more resourceful nodes to be willing to share their model updates, they need to be fairly incentivized. This paper explores an incentive design that guarantees fairness so that nodes receive rewards commensurate to their contributions. Our approach leverages an explore-then-exploit formulation to estimate the nodes’ contributions (i. e. , exploration) for realizing our theoretically guaranteed fair incentives (i. e. , exploitation). However, we observe a "rich get richer" phenomenon arising from the existing approaches to guarantee fairness and it discourages the participation of the less resourceful nodes. To remedy this, we additionally preserve asymptotic equality, i. e. , less resourceful nodes achieve equal performance eventually to the more resourceful/“rich” nodes. We empirically demonstrate in two settings with real-world streaming data: federated online incremental learning and federated reinforcement learning, that our proposed approach outperforms existing baselines in fairness and learning performance while remaining competitive in preserving equality.

TMLR Journal 2023 Journal Article

Mitigating Real-World Distribution Shifts in the Fourier Domain

• Kiran Krishnamachari
• See-Kiong Ng
• Chuan-Sheng Foo

While machine learning systems can be highly accurate in their training environments, their performance in real-world deployments can suffer significantly due to distribution shifts. Real-world distribution shifts involve various input distortions due to noise, weather, device and other variations. Many real-world distribution shifts are not represented in standard domain adaptation datasets and prior empirical work has shown that domain adaptation methods developed using these standard datasets may not generalize well to real-world distribution shifts. Furthermore, motivated by observations of the sensitivity of deep neural networks (DNN) to the spectral statistics of data, which can vary in real-world scenarios, we propose Fourier Moment Matching (FMM), a model-agnostic input transformation that matches the Fourier-amplitude statistics of source to target data using unlabeled samples. We demonstrate through extensive empirical evaluations across time-series, image classification and semantic segmentation tasks that FMM is effective both individually and when combined with a variety of existing methods to overcome real-world distribution shifts. Code is available at https://github.com/kiranchari/FourierMomentMatching.

ICML Conference 2022 Conference Paper

Efficient Distributionally Robust Bayesian Optimization with Worst-case Sensitivity

• Sebastian Shenghong Tay
• Chuan-Sheng Foo
• Daisuke Urano
• Richalynn Leong
• Bryan Kian Hsiang Low

In distributionally robust Bayesian optimization (DRBO), an exact computation of the worst-case expected value requires solving an expensive convex optimization problem. We develop a fast approximation of the worst-case expected value based on the notion of worst-case sensitivity that caters to arbitrary convex distribution distances. We provide a regret bound for our novel DRBO algorithm with the fast approximation, and empirically show it is competitive with that using the exact worst-case expected value while incurring significantly less computation time. In order to guide the choice of distribution distance to be used with DRBO, we show that our approximation implicitly optimizes an objective close to an interpretable risk-sensitive value.

IJCAI Conference 2022 Conference Paper

Few-Shot Adaptation of Pre-Trained Networks for Domain Shift

• Wenyu Zhang
• Li Shen
• Wanyue Zhang
• Chuan-Sheng Foo

Deep networks are prone to performance degradation when there is a domain shift between the source (training) data and target (test) data. Recent test-time adaptation methods update batch normalization layers of pre-trained source models deployed in new target environments with streaming data. Although these methods can adapt on-the-fly without first collecting a large target domain dataset, their performance is dependent on streaming conditions such as mini-batch size and class-distribution which can be unpredictable in practice. In this work, we propose a framework for few-shot domain adaptation to address the practical challenges of data-efficient adaptation. Specifically, we propose a constrained optimization of feature normalization statistics in pre-trained source models supervised by a small target domain support set. Our method is easy to implement and improves source model performance with as little as one sample per class for classification tasks. Extensive experiments on 5 cross-domain classification and 4 semantic segmentation datasets show that our proposed method achieves more accurate and reliable performance than test-time adaptation, while not being constrained by streaming conditions.

TMLR Journal 2022 Journal Article

Fourier Sensitivity and Regularization of Computer Vision Models

• Kiran Krishnamachari
• See-Kiong Ng
• Chuan-Sheng Foo

Recent work has empirically shown that deep neural networks latch on to the Fourier statistics of training data and show increased sensitivity to Fourier-basis directions in the input. Understanding and modifying this Fourier-sensitivity of computer vision models may help improve their robustness, hence, in this paper we study the frequency sensitivity characteristics of deep neural networks using a principled approach. We first propose a $\textbf{\textit{basis trick}}$, proving that unitary transformations of the input-gradient of a function can be used to compute its gradient in the basis induced by the transformation. Using this result, we propose a general measure of any differentiable computer vision model's $\textit{\textbf{Fourier-sensitivity}}$ using the unitary Fourier-transform of its input-gradient. When applied to deep neural networks, we find that computer vision models are consistently sensitive to particular frequencies dependent on the dataset, training method and architecture. Based on this measure, we further propose a $\textit{\textbf{Fourier-regularization}}$ framework to modify the Fourier-sensitivities and frequency bias of models. Using our proposed regularizer-family, we demonstrate that deep neural networks obtain improved classification accuracy on robustness evaluations.

ICLR Conference 2021 Conference Paper

ARMOURED: Adversarially Robust MOdels using Unlabeled data by REgularizing Diversity

• Kangkang Lu 0001
• Cuong Manh Nguyen
• Xun Xu 0002
• Kiran Chari
• Yu Jing Goh
• Chuan-Sheng Foo

Adversarial attacks pose a major challenge for modern deep neural networks. Recent advancements show that adversarially robust generalization requires a large amount of labeled data for training. If annotation becomes a burden, can unlabeled data help bridge the gap? In this paper, we propose ARMOURED, an adversarially robust training method based on semi-supervised learning that consists of two components. The first component applies multi-view learning to simultaneously optimize multiple independent networks and utilizes unlabeled data to enforce labeling consistency. The second component reduces adversarial transferability among the networks via diversity regularizers inspired by determinantal point processes and entropy maximization. Experimental results show that under small perturbation budgets, ARMOURED is robust against strong adaptive adversaries. Notably, ARMOURED does not rely on generating adversarial samples during training. When used in combination with adversarial training, ARMOURED yields competitive performance with the state-of-the-art adversarially-robust benchmarks on SVHN and outperforms them on CIFAR-10, while offering higher clean accuracy.

ICML Conference 2009 Conference Paper

A majorization-minimization algorithm for (multiple) hyperparameter learning

• Chuan-Sheng Foo
• Chuong B. Do
• Andrew Y. Ng

We present a general Bayesian framework for hyperparameter tuning in L 2 -regularized supervised learning models. Paradoxically, our algorithm works by first analytically integrating out the hyperparameters from the model. We find a local optimum of the resulting non-convex optimization problem efficiently using a majorization-minimization (MM) algorithm, in which the non-convex problem is reduced to a series of convex L 2 -regularized parameter estimation tasks. The principal appeal of our method is its simplicity: the updates for choosing the L 2 -regularized subproblems in each step are trivial to implement (or even perform by hand), and each subproblem can be efficiently solved by adapting existing solvers. Empirical results on a variety of supervised learning models show that our algorithm is competitive with both grid-search and gradient-based algorithms, but is more efficient and far easier to implement.

ICML Conference 2009 Conference Paper

Proximal regularization for online and batch learning

• Chuong B. Do
• Quoc V. Le
• Chuan-Sheng Foo

NeurIPS Conference 2007 Conference Paper

Efficient multiple hyperparameter learning for log-linear models

• Chuan-Sheng Foo
• Chuong B.
• Andrew Ng

Using multiple regularization hyperparameters is an effective method for managing model complexity in problems where input features have varying amounts of noise. While algorithms for choosing multiple hyperparameters are often used in neural networks and support vector machines, they are not common in structured prediction tasks, such as sequence labeling or parsing. In this paper, we consider the problem of learning regularization hyperparameters for log-linear models, a class of probabilistic models for structured prediction tasks which includes conditional random fields (CRFs). Using an implicit differentiation trick, we derive an efficient gradient-based method for learning Gaussian regularization priors with multiple hyperparameters. In both simulations and the real-world task of computational RNA secondary structure prediction, we find that multiple hyperparameter learning provides a significant boost in accuracy compared to models learned using only a single regularization hyperparameter.