-
Cross-Cluster Shifting for Efficient and Effective 3D Object Detection in Autonomous Driving
Authors:
Zhili Chen,
Kien T. Pham,
Maosheng Ye,
Zhiqiang Shen,
Qifeng Chen
Abstract:
We present a new 3D point-based detector model, named Shift-SSD, for precise 3D object detection in autonomous driving. Traditional point-based 3D object detectors often employ architectures that rely on a progressive downsampling of points. While this method effectively reduces computational demands and increases receptive fields, it will compromise the preservation of crucial non-local informati…
▽ More
We present a new 3D point-based detector model, named Shift-SSD, for precise 3D object detection in autonomous driving. Traditional point-based 3D object detectors often employ architectures that rely on a progressive downsampling of points. While this method effectively reduces computational demands and increases receptive fields, it will compromise the preservation of crucial non-local information for accurate 3D object detection, especially in the complex driving scenarios. To address this, we introduce an intriguing Cross-Cluster Shifting operation to unleash the representation capacity of the point-based detector by efficiently modeling longer-range inter-dependency while including only a negligible overhead. Concretely, the Cross-Cluster Shifting operation enhances the conventional design by shifting partial channels from neighboring clusters, which enables richer interaction with non-local regions and thus enlarges the receptive field of clusters. We conduct extensive experiments on the KITTI, Waymo, and nuScenes datasets, and the results demonstrate the state-of-the-art performance of Shift-SSD in both detection accuracy and runtime efficiency.
△ Less
Submitted 10 March, 2024;
originally announced March 2024.
-
A Schrödinger Equation for Evolutionary Dynamics
Authors:
Vi D. Ao,
Duy V. Tran,
Kien T. Pham,
Duc M. Nguyen,
Huy D. Tran,
Tuan K. Do,
Van H. Do,
Trung V. Phan
Abstract:
We establish an analogy between the Fokker-Planck equation describing evolutionary landscape dynamics and the Schrödinger equation which characterizes quantum mechanical particles, showing how a population with multiple genetic traits evolves analogously to a wavefunction under a multi-dimensional energy potential in imaginary time. Furthermore, we discover within this analogy that the stationary…
▽ More
We establish an analogy between the Fokker-Planck equation describing evolutionary landscape dynamics and the Schrödinger equation which characterizes quantum mechanical particles, showing how a population with multiple genetic traits evolves analogously to a wavefunction under a multi-dimensional energy potential in imaginary time. Furthermore, we discover within this analogy that the stationary population distribution on the landscape corresponds exactly to the ground-state wavefunction. This mathematical equivalence grants entry to a wide range of analytical tools developed by the quantum mechanics community, such as the Rayleigh-Ritz variational method and the Rayleigh-Schrödinger perturbation theory, allowing us to not only make reasonable quantitative assessments but also explore fundamental biological inquiries. We demonstrate the effectiveness of these tools by estimating the population success on landscapes where precise answers are elusive, and unveiling the ecological consequences of stress-induced mutagenesis -- a prevalent evolutionary mechanism in pathogenic and neoplastic systems. We show that, even in a unchanging environment, a sharp mutational burst resulting from stress can always be advantageous, while a gradual increase only enhances population size when the number of relevant evolving traits is limited. Our interdisciplinary approach offers novel insights, opening up new avenues for deeper understanding and predictive capability regarding the complex dynamics of evolving populations.
△ Less
Submitted 31 August, 2023; v1 submitted 29 July, 2023;
originally announced July 2023.
-
Abstractive Text Summarization Using the BRIO Training Paradigm
Authors:
Khang Nhut Lam,
Thieu Gia Doan,
Khang Thua Pham,
Jugal Kalita
Abstract:
Summary sentences produced by abstractive summarization models may be coherent and comprehensive, but they lack control and rely heavily on reference summaries. The BRIO training paradigm assumes a non-deterministic distribution to reduce the model's dependence on reference summaries, and improve model performance during inference. This paper presents a straightforward but effective technique to i…
▽ More
Summary sentences produced by abstractive summarization models may be coherent and comprehensive, but they lack control and rely heavily on reference summaries. The BRIO training paradigm assumes a non-deterministic distribution to reduce the model's dependence on reference summaries, and improve model performance during inference. This paper presents a straightforward but effective technique to improve abstractive summaries by fine-tuning pre-trained language models, and training them with the BRIO paradigm. We build a text summarization dataset for Vietnamese, called VieSum. We perform experiments with abstractive summarization models trained with the BRIO paradigm on the CNNDM and the VieSum datasets. The results show that the models, trained on basic hardware, outperform all existing abstractive summarization models, especially for Vietnamese.
△ Less
Submitted 23 May, 2023;
originally announced May 2023.
-
Stress-Induced Mutagenesis Can Further Boost Population Success in Static Ecology
Authors:
Kien T. Pham,
Duc M. Nguyen,
Duy V. Tran,
Vi D. Ao,
Huy D. Tran,
Tuan K. Do,
Trung V. Phan
Abstract:
We have developed a mathematical model that captures stress-induced mutagenesis, a fundamental aspect of pathogenic and neoplastic evolutionary dynamics, on the fitness landscape with multiple relevant genetic traits as a high-dimensional Euclidean space. In this framework, stress-induced mutagenesis manifests as a heterogeneous diffusion process. We show how increasing mutations, and thus reducin…
▽ More
We have developed a mathematical model that captures stress-induced mutagenesis, a fundamental aspect of pathogenic and neoplastic evolutionary dynamics, on the fitness landscape with multiple relevant genetic traits as a high-dimensional Euclidean space. In this framework, stress-induced mutagenesis manifests as a heterogeneous diffusion process. We show how increasing mutations, and thus reducing exploitation, in a static ecology with fixed carrying capacity and maximum growth rates, can paradoxically boost population size. Remarkably, this unexpected biophysical phenomenon applies universally to any number of traits.
△ Less
Submitted 16 March, 2023;
originally announced March 2023.
