Information Theory and Reliable Communication

Most machine-learning methods focus on classifying instances whose classes have already been seen in training. In practice, many applications require classifying instances whose classes have not been seen previously. Zero-shot learning is a powerful and promising learning paradigm, in which the classes covered by training instances and the classes we aim to classify are disjoint. In this paper, we provide a comprehensive survey of zero-shot learning. First of all, we provide an overview of zero-shot learning. According to the data utilized in model optimization, we classify zero-shot learning into three learning settings. Second, we describe different semantic spaces adopted in existing zero-shot learning works. Third, we categorize existing zero-shot learning methods and introduce representative methods under each category. Fourth, we discuss different applications of zero-shot learning. Finally, we highlight promising future research directions of zero-shot learning.

A Survey of Zero-Shot Learning: Settings, Methods, and Applications

Driven by the need to cope with exponentially growing mobile data traffic and to support new traffic types from massive numbers of machine-type devices, academia and industry are thinking beyond the current generation of mobile cellular networks to chalk a path towards fifth generation (5G) mobile networks. Several new approaches and technologies are being considered as potential elements making up such a future mobile network, including cloud RANs, application of SDN principles, exploiting new and unused portions of spectrum, use of massive MIMO and full-duplex communications. Research on these technologies requires realistic and flexible experimentation platforms that offer a wide range of experimentation modes from real-world experimentation to controlled and scalable evaluations while at the same time retaining backward compatibility with current generation systems. Towards this end, we present OpenAirInterface (OAI) as a suitably flexible platform. In addition, we discuss the use of OAI in the context of several widely mentioned 5G research directions.

/pdf/openairinterface-a-flexible-platform-for-5g-research-2sc2xt49j2.pdf

OpenAirInterface: A Flexible Platform for 5G Research

Device-free passive indoor localization is playing a critical role in many applications such as elderly care, intrusion detection, smart home, etc. However, existing device-free localization systems either suffer from labor-intensive offline training or require dedicated special-purpose devices. To address the challenges, we present our system named MaTrack, which is implemented on commodity off-the-shelf Intel 5300 Wi-Fi cards. MaTrack proposes a novel Dynamic-MUSIC method to detect the subtle reflection signals from human body and further differentiate them from those reflected signals from static objects (furniture, walls, etc.) to identify the human target's angle for localization. MaTrack does not require any offline training compared to existing signature-based systems and is insensitive to changes in environment. With just two receivers, MaTrack is able to achieve a median localization accuracy below 0.6 m when the human is walking, outperforming the state-of-the-art schemes.

https://ink.library.smu.edu.sg/cgi/viewcontent.cgi?article=4391&context=sis_research

Dynamic-MUSIC: accurate device-free indoor localization

Widely regarded as one of the most promising emerging technologies for driving the future development of wireless communications, cognitive radio has the potential to mitigate the problem of increasing radio spectrum scarcity through dynamic spectrum allocation. Drawing on fundamental elements of information theory, network theory, propagation, optimisation and signal processing, a team of leading experts present a systematic treatment of the core physical and networking principles of cognitive radio and explore key design considerations for the development of new cognitive radio systems. Containing all the underlying principles you need to develop practical applications in cognitive radio, this book is an essential reference for students, researchers and practitioners alike in the field of wireless communications and signal processing.

/pdf/principles-of-cognitive-radio-2cmaz71gzs.pdf

Principles Of Cognitive Radio

This paper presents the design ofWARP, a custom platform for research in advanced wireless algorithms and applications. The platform consists of both custom hardware and FPGA implementations of key communications blocks. The hardware consists of FPGA-based processing boards coupled to wideband radios and other I/O interfaces; the algorithm implementations already include a flexible OFDM physical layer. Both the hardware specifications and algorithm implementations will be freely available to academic researchers to enable the development of a widely disseminated, highly capable platform for wireless research.

Design of WARP: A wireless open-access research platform

In this paper, we introduce the wireless open-access research platform (WARP) developed at CMC lab, Rice University. WARP provides a scalable and configurable platform mainly designed to prototype wireless communication algorithms for educational and research oriented applications. Its programmability and flexibility makes it easy to implement various physical and network layer protocols and standards. Moreover, the online open-access WARP repository is used to document and share different wireless architectures and cross-layer designs developed at educational and research centers. This repository is a fast and easy solution for students and researchers with a wide range of backgrounds in hardware implementation and algorithm development to collaborate and initiate multi-disciplinary system designs.

/pdf/warp-a-unified-wireless-network-testbed-for-education-and-5gtfoq6mmy.pdf

WARP, a Unified Wireless Network Testbed for Education and Research

Fine-grained image classification, which targets at distinguishing subtle distinctions among various subordinate categories, remains a very difficult task due to the high annotation cost of enormous fine-grained categories. To cope with the scarcity of well-labeled training images, existing works mainly follow two research directions: 1) utilize freely available web images without human annotation; 2) only annotate some fine-grained categories and transfer the knowledge to other fine-grained categories, which falls into the scope of zero-shot learning (ZSL). However, the above two directions have their own drawbacks. For the first direction, the labels of web images are very noisy and the data distribution between web images and test images are considerably different. For the second direction, the performance gap between ZSL and traditional supervised learning is still very large. The drawbacks of the above two directions motivate us to design a new framework which can jointly leverage both web data and auxiliary labeled categories to predict the test categories that are not associated with any well-labeled training images. Comprehensive experiments on three benchmark datasets demonstrate the effectiveness of our proposed framework.

https://openaccess.thecvf.com/content_cvpr_2018/papers/Niu_Webly_Supervised_Learning_CVPR_2018_paper.pdf

Webly Supervised Learning Meets Zero-shot Learning: A Hybrid Approach for Fine-Grained Classification

The goal of this article is to discuss the implementation of two cooperative protocols - decode-and-forward (DF) and estimate-and-forward (EF). In each case, the starting point is an information theoretic random coding scheme, which motivates a practical code construction. Low-density parity check (LDPC) codes are used as components in both protocols. In DF relaying, the problem boils down to a search for multi-rate LDPC codes that simultaneously perform well for the source-relay and the source- destination links. On the other hand, in EF relaying, the challenge is to design an efficient quantizer for forming a compressed estimate at the relay. The performance of each scheme is comparable to theoretical limits.

Code designs for cooperative communication

As tons of photos are being uploaded to public websites (e.g., Flickr, Bing, and Google) every day, learning from web data has become an increasingly popular research direction because of freely available web resources, which is also referred to as webly supervised learning. Nevertheless, the performance gap between webly supervised learning and traditional supervised learning is still very large, owning to the label noise of web data. To be exact, the labels of images crawled from public websites are very noisy and often inaccurate. Some existing works tend to facilitate learning from web data with the aid of extra information, such as augmenting or purifying web data by virtue of instance-level supervision, which is usually in demand of heavy manual annotation. Instead, we propose to tackle the label noise by leveraging more accessible category-level supervision. In particular, we build our method upon variational autoencoder (VAE), in which the classification network is attached on the hidden layer of VAE in a way that the classification network and VAE can jointly leverage the category-level hybrid semantic information. The effectiveness of our proposed method is clearly demonstrated by extensive experiments on three benchmark datasets.

A. Sabharwal

Papers

Design of WARP: A wireless open-access research platform

WARP, a Unified Wireless Network Testbed for Education and Research

Webly Supervised Learning Meets Zero-shot Learning: A Hybrid Approach for Fine-Grained Classification

Code designs for cooperative communication

Learning from Noisy Web Data with Category-level Supervision