Lingkai Kong

Lingkai Kong

Auto-generated profile, based on publicly available information

Lingkai Kong

Georgia Institute of Technol- ogy

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 January 2022
Verified
Conference Paper
Open access
PDF file
YYYue Yu
Yue Yu
LKLingkai Kong
Lingkai Kong
JZJieyu Zhang
Jieyu Zhang
RZRongzhi Zhang
Rongzhi Zhang
CZChao Zhang
Chao Zhang
Yue Yu, Lingkai Kong, Jieyu Zhang, Rongzhi Zhang, Chao Zhang. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

4 August 2023
Verified
Conference Paper
Open access
YZYuchen Zhuang
Yuchen Zhuang
YYYue Yu
Yue Yu
LKLingkai Kong
Lingkai Kong
XCXiang Chen
Xiang Chen
CZChao Zhang
Chao Zhang
Learning from noisy labels is a challenge that arises in many real-world applications where training data can contain incorrect or corrupted labels. When fine-tuning language models with noisy labels, models can easily overfit the label noise, leading to decreased performance. Most existing methods for learning from noisy labels use static input features for denoising, but these methods are limited by the information they can provide on true label distributions and can result in biased or incorrect predictions. In this work, we propose the Dynamics-Enhanced Generative Model (DyGen), which uses dynamic patterns in the embedding space during the fine-tuning process of language models to improve noisy label predictions. DyGen uses the variational auto-encoding framework to infer the posterior distributions of true labels from noisy labels and training dynamics. Additionally, a co-regularization mechanism is used to minimize the impact of potentially noisy labels and priors. DyGen demonstrates an average accuracy improvement of 3.10% on two synthetic noise datasets and 1.48% on three real-world noise datasets compared to the previous state-of-the-art. Extensive experiments and analyses show the effectiveness of each component in DyGen. Our code is available for reproducibility on GitHub.

4 August 2023
Verified
Conference Paper
Open access
LKLingkai Kong
Lingkai Kong
HKHarshavardhan Kamarthi
Harshavardhan Kamarthi
PCPeng Chen
Peng Chen
BPB. Aditya Prakash
B. Aditya Prakash
CZChao Zhang
Chao Zhang
Deep neural networks (DNNs) have achieved enormous success in a wide range of domains, such as computer vision, natural language processing and scientific areas. However, one key bottleneck of DNNs is that they are ignorant about the uncertainties in their predictions. They can produce wildly wrong predictions without realizing, and can even be confident about their mistakes. Such mistakes can cause misguided decisions-sometimes catastrophic in critical applications, ranging from self-driving cars to cyber security to automatic medical diagnosis. In this tutorial, we present recent advancements in uncertainty quantification for DNNs and their applications across various domains. We first provide an overview of the motivation behind uncertainty quantification, different sources of uncertainty, and evaluation metrics. Then, we delve into several representative uncertainty quantification methods for predictive models, including ensembles, Bayesian neural networks, conformal prediction, and others. We go on to discuss how uncertainty can be utilized for label-efficient learning, continual learning, robust decision-making, and experimental design. Furthermore, we showcase examples of uncertainty-aware DNNs in various domains, such as health, robotics, and scientific machine learning. Finally, we summarize open challenges and future directions in this area.

22 June 2021 in Journal of Lightwave Technology
Verified
Journal Article
Closed access
SYShuang Yao
Shuang Yao
CHChin-Wei Hsu
Chin-Wei Hsu
LKLingkai Kong
Lingkai Kong
QZQi Zhou
Qi Zhou
SSShuyi Shen
Shuyi Shen
GCGee-Kung Chang
Gee-Kung Chang
Quality of Transmission (QoT) estimation, where the received signal quality is predicted before deployment, plays a significant role in efficient resource utilization, such as determining the optimal transmission configuration. Traditionally, it is implemented with analytical models, and often accompanied by large link margins to account for the low estimation accuracy. Machine learning (ML) based methods have been recently demonstrated as an alternative solution with high accuracy. However, they require a large number of training data, which is often expensive to obtain in the context of QoT estimation. In this paper, we use active learning (AL) to achieve data efficient QoT estimation. A learner actively selects the training data to be labeled by applying the strategy of uncertainty sampling which favors data with high model uncertainty. A data selection algorithm compatible with the widely studied artificial neural network (ANN)-based QoT estimator is proposed and experimentally demonstrated in a fiber-wireless integrated testbed. Monte Carlo dropout (MC dropout) is utilized to calculate model uncertainty. To achieve a mean squared error (MSE) of 0.055, the number of training data can be reduced by more than 25% compared with the conventional passive ML. The algorithm is also investigated under different sampling settings and the impact of hyperparameters is discussed.

4 August 2023
Verified
Conference Paper
Open access
HKHarshavardhan Kamarthi
Harshavardhan Kamarthi
LKLingkai Kong
Lingkai Kong
ARAlexander Rodriguez
Alexander Rodriguez
CZChao Zhang
Chao Zhang
BPB. Aditya Prakash
B. Aditya Prakash
Probabilistic hierarchical time-series forecasting is an important variant of time-series forecasting, where the goal is to model and forecast multivariate time-series that have hierarchical relations. Previous works assume rigid consistency over the given hierarchies and do not adapt well to real-world data that show deviation from this assumption. Moreover, recent state-of-art neural probabilistic methods also impose hierarchical relations on point predictions and samples of the predictive distribution. This does not account for full forecast distributions being consistent with the hierarchy and leading to poorly calibrated forecasts. We close both these gaps and propose PROFHiT, a probabilistic hierarchical forecasting model that jointly models forecast distributions over the entire hierarchy. PROFHiT (1) uses a flexible probabilistic Bayesian approach and (2) introduces soft distributional consistency regularization that enables end-to-end learning of the entire forecast distribution leveraging information from the underlying hierarchy. This enables calibrated forecasts as well as adaptation to real-life data with varied hierarchical consistency. PROFHiT provides 41-88% better performance in accuracy and significantly better calibration over a wide range of dataset consistency. Furthermore, PROFHiT adapts to missing data and can provide reliable forecasts even if up to 10% of input time-series data is missing, whereas other methods' performance severely degrades by over 70%

1 November 2014
Verified
Conference Paper
Closed access
NKNai-Chung Kuo
Nai-Chung Kuo
BYBonjern Yang
Bonjern Yang
CWChaoying Wu
Chaoying Wu
LKLingkai Kong
Lingkai Kong
AWAngie Wang
Angie Wang
BNBorivoje Nikolic
Borivoje Nikolic
This paper demonstrates a CMOS digital polar transmitter with flip-chip interconnection to low-temperature co-fired ceramic (LTCC) interposers. The LTCC interposers contain the PA output balun targeting different operating frequency bands, and the reconfiguration in the carrier frequency is achieved by selecting an appropriate LTCC interposer. The same CMOS core transmitter is reused for different frequency bands. In this design, an output power higher than 22 dBm from 0.6 to 2.4 GHz is demonstrated, with peak power of 27.1 dBm and peak efficiency of 52%. The polar transmitter includes 9-bit phase interpolation and 8-bit amplitude modulation, suitable and verified as a multi-standard universal digital modulator.