Author name cluster

Cheng-Hao Tu

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

1 author row

NeurIPS Conference 2024 Conference Paper

Fine-Tuning is Fine, if Calibrated

Zheda Mai
Arpita Chowdhury
Ping Zhang
Cheng-Hao Tu
Hong-You Chen
Vardaan Pahuja
Tanya Berger-Wolf
Song Gao

Fine-tuning is arguably the most straightforward way to tailor a pre-trained model (e. g. , a foundation model) to downstream applications, but it also comes with the risk of losing valuable knowledge the model had learned in pre-training. For example, fine-tuning a pre-trained classifier capable of recognizing a large number of classes to master a subset of classes at hand is shown to drastically degrade the model's accuracy in the other classes it had previously learned. As such, it is hard to further use the fine-tuned model when it encounters classes beyond the fine-tuning data. In this paper, we systematically dissect the issue, aiming to answer the fundamental question, "What has been damaged in the fine-tuned model? " To our surprise, we find that the fine-tuned model neither forgets the relationship among the other classes nor degrades the features to recognize these classes. Instead, the fine-tuned model often produces more discriminative features for these other classes, even if they were missing during fine-tuning! What really hurts the accuracy is the discrepant logit scales between the fine-tuning classes and the other classes, implying that a simple post-processing calibration would bring back the pre-trained model's capability and at the same time unveil the feature improvement over all classes. We conduct an extensive empirical study to demonstrate the robustness of our findings and provide preliminary explanations underlying them, suggesting new directions for future theoretical analysis.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

Holistic Transfer: Towards Non-Disruptive Fine-Tuning with Partial Target Data

Cheng-Hao Tu
Hong-You Chen
Zheda Mai
Jike Zhong
Vardaan Pahuja
Tanya Berger-Wolf
Song Gao
Charles Stewart

We propose a learning problem involving adapting a pre-trained source model to the target domain for classifying all classes that appeared in the source data, using target data that covers only a partial label space. This problem is practical, as it is unrealistic for the target end-users to collect data for all classes prior to adaptation. However, it has received limited attention in the literature. To shed light on this issue, we construct benchmark datasets and conduct extensive experiments to uncover the inherent challenges. We found a dilemma --- on the one hand, adapting to the new target domain is important to claim better performance; on the other hand, we observe that preserving the classification accuracy of classes missing in the target adaptation data is highly challenging, let alone improving them. To tackle this, we identify two key directions: 1) disentangling domain gradients from classification gradients, and 2) preserving class relationships. We present several effective solutions that maintain the accuracy of the missing classes and enhance the overall performance, establishing solid baselines for holistic transfer of pre-trained models with partial target data.

PDF Details

AAAI Conference 2023 Conference Paper

Learning Fractals by Gradient Descent

Cheng-Hao Tu
Hong-You Chen
David Carlyn
Wei-Lun Chao

Fractals are geometric shapes that can display complex and self-similar patterns found in nature (e.g., clouds and plants). Recent works in visual recognition have leveraged this property to create random fractal images for model pre-training. In this paper, we study the inverse problem --- given a target image (not necessarily a fractal), we aim to generate a fractal image that looks like it. We propose a novel approach that learns the parameters underlying a fractal image via gradient descent. We show that our approach can find fractal parameters of high visual quality and be compatible with different loss functions, opening up several potentials, e.g., learning fractals for downstream tasks, scientific understanding, etc.

PDF Details DOI

NeurIPS Conference 2019 Conference Paper

Compacting, Picking and Growing for Unforgetting Continual Learning

Ching-Yi Hung
Cheng-Hao Tu
Cheng-En Wu
Chien-Hung Chen
Yi-Ming Chan
Chu-Song Chen

Continual lifelong learning is essential to many applications. In this paper, we propose a simple but effective approach to continual deep learning. Our approach leverages the principles of deep model compression, critical weights selection, and progressive networks expansion. By enforcing their integration in an iterative manner, we introduce an incremental learning method that is scalable to the number of sequential tasks in a continual learning process. Our approach is easy to implement and owns several favorable characteristics. First, it can avoid forgetting (i. e. , learn new tasks while remembering all previous tasks). Second, it allows model expansion but can maintain the model compactness when handling sequential tasks. Besides, through our compaction and selection/expansion mechanism, we show that the knowledge accumulated through learning previous tasks is helpful to build a better model for the new tasks compared to training the models independently with tasks. Experimental results show that our approach can incrementally learn a deep model tackling multiple tasks without forgetting, while the model compactness is maintained with the performance more satisfiable than individual task training.

PDF Details

YNIMG Journal 2009 Journal Article

Abnormal cerebral metabolism during menstrual pain in primary dysmenorrhea

Cheng-Hao Tu
David M. Niddam
Hsiang-Tai Chao
Ren-Shyan Liu
Ren-Jen Hwang
Tzu-Chen Yeh
Jen-Chuen Hsieh

Primary dysmenorrhea (PDM, menstrual pain without pelvic abnormality) is the most common gynecological disorder for women in the reproductive age. It is characterized by cramping pain and enhanced pain sensitivity during the menstruation period. PDM has been associated with peripheral and central sensitization. Abnormal brain mechanisms may further contribute to development and maintenance of the state. Using fluoro-deoxyglucose positron emission tomography, increased activity was observed in prefrontal/orbitofrontal regions and left ventral posterior thalamus while decreased activity mainly was observed in sensorimotor regions of the left hemisphere at onset compared to offset of PDM. These results were specific to menstrual pain and were not found in menstrual matched controls. Orbitofrontal activities were positively related to while somatosensory activities where negatively related to subjective pain ratings. These results show that ongoing menstrual pain in PDM is accompanied by abnormal brain metabolism. Disinhibition of thalamo-orbitofrontal-prefrontal networks may contribute to the generation of pain and hyperalgesia in PDM possibly by maintaining spinal and thalamic sensitization while increasing negative affect. Excessive excitatory input during menstrual pain may induce compensatory inhibitory mechanism in several somatic sensorimotor regions.

Details DOI