Source author record

Tianyi Hu

Tianyi Hu appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Machine Learning Artificial Intelligence Computer Vision cond-mat.mes-hall Information Retrieval

Catalog footprint

What is connected

3works

5topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Content-Based Landmark Retrieval Combining Global and Local Features using Siamese Neural Networks

In this work, we present a method for landmark retrieval that utilizes global and local features. A Siamese network is used for global feature extraction and metric learning, which gives an initial ranking of the landmark search. We utilize the extracted feature maps from the Siamese architecture as local descriptors, the search results are then further refined using a cosine similarity between local descriptors. We conduct a deeper analysis of the Google Landmark Dataset, which is used for evaluation, and augment the dataset to handle various intra-class variances. Furthermore, we conduct several experiments to compare the effects of transfer learning and metric learning, as well as experiments using other local descriptors. We show that a re-ranking using local features can improve the search results. We believe that the proposed local feature extraction using cosine similarity is a simple approach that can be extended to many other retrieval tasks.

preprint2022arXiv

Hybrid Learning for Orchestrating Deep Learning Inference in Multi-user Edge-cloud Networks

Deep-learning-based intelligent services have become prevalent in cyber-physical applications including smart cities and health-care. Collaborative end-edge-cloud computing for deep learning provides a range of performance and efficiency that can address application requirements through computation offloading. The decision to offload computation is a communication-computation co-optimization problem that varies with both system parameters (e.g., network condition) and workload characteristics (e.g., inputs). Identifying optimal orchestration considering the cross-layer opportunities and requirements in the face of varying system dynamics is a challenging multi-dimensional problem. While Reinforcement Learning (RL) approaches have been proposed earlier, they suffer from a large number of trial-and-errors during the learning process resulting in excessive time and resource consumption. We present a Hybrid Learning orchestration framework that reduces the number of interactions with the system environment by combining model-based and model-free reinforcement learning. Our Deep Learning inference orchestration strategy employs reinforcement learning to find the optimal orchestration policy. Furthermore, we deploy Hybrid Learning (HL) to accelerate the RL learning process and reduce the number of direct samplings. We demonstrate efficacy of our HL strategy through experimental comparison with state-of-the-art RL-based inference orchestration, demonstrating that our HL strategy accelerates the learning process by up to 166.6x.

preprint2021arXiv

Kekule Lattice in Graphdiyne: Coexistence of Phononic and Electronic Higher-Order Band Topology

The topological physics has been extensively studied in different kinds of bosonic and fermionic systems, ranging from artificial structures to natural materials. However, the coexistence of topological phonon and electron in one single material is seldom reported. Recently, graphdiyne is proposed to be a two-dimensional (2D) electronic second-order topological insulator (SOTI). In this work, based on density-functional tight-binding calculations, we found that graphdiyne is equivalent to the Kekule lattice, also realizing a 2D phononic SOTI in both out-of-plane and in-plane modes. Depending on edge terminations, the characterized topological corner states can be either inside or outside the bulk gap, which are tunable by local corner potential. Most remarkably, a unique selectivity of space and symmetry is revealed in electron-phonon coupling between the localized phononic and electronic topological corner states. Our results not only demonstrate the phononic higher-order band topology in a real carbon material, but also provide an opportunity to investigate the interplay between phononic and electronic higher-order topological states.