News

We are so happy to welcome a new MCL member, Xuechun Hua joining MCL this semester. Here is a quick interview with Xuechun:

1. Could you briefly introduce yourself and your research interests?

My name is Xuechun Hua, I am a second year graduate student at Viterbi pursuing a Computer Science master degree. I finished my undergraduate study in Nanjing University(NJU) and focused on game theory. I developed my research interest in interpretable machine learning and computer vision.

What is your impression about MCL and USC?

USC has offered me an exceptional environment that fosters both my personal and academic development. The atmosphere at MCL is one of serious academic pursuit combined with a strong sense of collaboration. I am deeply thankful for the extensive support I’ve received from senior members of MCL. Professor Kuo stands out as an endlessly energetic researcher, who greatly inspires us with his guidance and enthusiasm for exploration.

3. What is your future expectation and plan in MCL?

I will be collaborating with Qingyang on a project focused on mesh reconstruction from point clouds, under the mentorship of Professor Kuo. Moving forward, I aim to enhance my research skills within MCL and contribute to advancements in 3D representation and greening learning.

In the world of digital images, turning raw sensor data into colorful pictures that we can see is a complex process. Demosaicing, a crucial step in this process, helps convert data from sensors into full-color images. It all started with Bayer color filter arrays, named after Bryce Bayer, which are grids of tiny sensors, each covered by a color filter—usually red, green, or blue.

But making this conversion isn’t easy. Real-world challenges like sensor noise and blurry motion can mess up the final image. And getting accurate data for training computers to do this job is time-consuming.

Because processing images can be slow, especially on devices like phones, we’re looking into simpler methods that still give good results. Recently we’re experimenting with a new regressor named “LQBoost”

LQBoost operates by conducting least-square regressions in successive iterations, gradually narrowing down the gap between its predictions and the actual targets. By focusing on minimizing the residuals—differences between predicted and actual values—LQBoost iteratively enhances its accuracy. Additionally, it employs a thresholding process to prune samples with residuals approaching zero, streamlining the regression process.

Taking LQBoost a step further, we integrate Local Neighborhood Transformation (LNT) to enrich the feature set and capture intricate data structures more effectively. This integration allows for a more nuanced understanding of the data, leading to improved predictions.

Before applying LQBoost to our demosaicing task, we perform a crucial preprocessing step. By clustering the dataset and utilizing cluster purity, we initialize the regression process effectively. This step ensures that each cluster receives an accurate initial prediction, setting the stage for LQBoost to refine these predictions through iterative regression.

Our goal is to create a demosaicing model that’s both accurate and fast. We’ve tested it thoroughly using standard image datasets, making [...]

In a Machine Learning framework, there are three fundamental components that play a crucial role: feature extraction, feature selection, and decision model. In the context of Green Learning, we have Saab transform and Least-squares Normal Transform for feature extraction. Regarding feature selection, we have the Discriminant Feature Test (DFT) and Relevant Feature Test (RFT). However, we do not have a green and interpretable solution for the decision model. For a long time, we applied gradient-boosted trees such as XGBoost or LightGBM as the classifier. Yet, it is known that XGBoost or LightGBM models sacrifice interpretability for performance. Also, the large model size of XGBoost or LightGBM is becoming a huge burden for Green Learning. Therefore, we are motivated to develop a green and interpretable classifier called SLMBoost. The idea is to train a boosting model with the Subspace Learning Machine (SLM). 

Let’s start by looking at a single SLM. In each SLM, we will first identify a discriminant subspace by a series of feature selection techniques, including DFT, RFT, removing correlated features, etc. Then, each SLM will learn a linear regression model on the selected subspace. Figure 1 illustrates a single SLM. To sum up, a single SLM is a linear least square model that operates on a subspace.

Further, we ensemble SLMs in a boosting fashion, which involves training a sequence of SLMs. In this approach, an SLM is trained to correct the errors made by the previous SLMs. To achieve this, the training target of an SLM is the residual of the accumulated result from all the SLMs before it. Two key factors to make this boosting work are changing different features and data subsets. By using DFT/RFT, we can zoom in on different [...]

Congratulations to Yifan Wang for passing his defense today. Yifan’s thesis is titled “Advanced Techniques for Green Image Coding via Hierarchical Vector Quantization.” His Dissertation Committee includes Jay Kuo (Chair), Antonio Ortega, and Aiichiro Nakano (Outside Member). The Committee members appreciated Yifan’s contributions to this field in solving a very challenging problem. MCL News team invited Yifan for a short talk on his thesis and PhD experience, and here is the summary. We thank Yifan for his kind sharing, and wish him all the best in the next journey. A high-level abstract of Yifan’s thesis is given below:

”We developed an image compression codec, “Green Image Codec,” which differs from the traditional intra-prediction/transform-based or deep learning-based codec. We aim to have a codec with low decoding complexity and reasonable performance. We utilize the multi-grid representation as the foundation for our vector quantization-based codec. We proposed several coding tools that were designed and specialized for VQ to improve the rate-distortion gain and reduce the coding complexity. Systemic RDO, which was missing from the traditional VQ-based codec, was added to our framework and relieved the burden of finding suitable coding parameters. We can perform better at a low bit rate compared to H.264 intra with similar complexity.”

Yifan shared his Ph.D. experience at MCL as follows :

“During my time at MCL, I am very grateful for Professor Kuo’s push, hard work, and wisdom; I learned a lot from him. Even though the Ph.D. life for my first year was struggling, the new research topic for VQ-based image coding with few previous references brought a lot of trouble. We made many trials and errors, and I almost lost confidence and motivation because of the difficulty of developing a new codec. However, Professor Kuo kept working hard and provided [...]

Clear cell renal cell carcinoma (ccRCC) is the most common type of kidney cancer. Programmed Death-Ligand 1 (PD-L1) is an important prognostic marker for ccRCC and CT radiomics can be a non-invasive modality to predict PD-L1.

In a previous study by Shieh et al. [1], the research team explored using radiomic features from computer tomography (CT) imaging to predict the TIME measurements via mIHC analysis. Tumor specimens were categorized as positive or negative, guided by varying thresholds of PD-L1 expression of the tumor epithelium. A total of 1,708 radiomic features were extracted from CT images using a custom-built radiomics pipeline. To predict tumor binary classification based on these radiomic features.

Building upon this work [1], our study explored a novel approach known as Green Learning (GL).

As shown in Fig. 1, our study employs a modularized solution that includes the following modules: the DFT module for feature dimension reduction and the LNT module for new feature generation. Fig [2] presents the sorted DFT loss curve. The x-axis represents the sorted dimension index, while the y-axis represents the corresponding DFT loss value. The colored points scattered on the curve denote the DFT loss of the top 10 variables of importance (VOI), which are identified from the prior study [1]. All these top 10 VOIs possess relatively low DFT loss. The DFT loss value of LNT feature is the lowest among the raw features.

We observed a significant improvement in the radiomics prediction performance of tumor epithelium PD-L1 > 1%, >5% and >10%. Compared to prior research, the AUROC values improved from 0.61 to 0.76, 0.75 to 0.85 and 0.85 to 0.88, respectively.

[1] A. T.-W. Shieh, S. Y. Cen, B. Varghese, D. H. Hwang, X. Lei, K. Gurumurthy, I. Siddiqi, [...]