Cross-Layer Adaptive Bitrate Streaming

Total Page:16

File Type:pdf, Size:1020Kb

Cross-Layer Adaptive Bitrate Streaming Research Collection Master Thesis Cross-layer Adaptive Bitrate Streaming Author(s): Dan, Alexandru Publication Date: 2020 Permanent Link: https://doi.org/10.3929/ethz-b-000443424 Rights / License: In Copyright - Non-Commercial Use Permitted This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use. ETH Library Cross-layer Adaptive Bitrate Streaming Master Thesis Alexandru Dan 12 September, 2020 Advisors: Prof. Dr. Ankit Singla, Melissa Licciardello Department of Computer Science, ETH Z¨urich Abstract Video traffic makes up most of today’s internet traffic. Yet, most trans- port layer protocols are oblivious to the nature of the traffic transmit- ted through them. Videos are divided in chunks and streamed at the application-level over HTTP connections. However, the underlying transport can not tell if the user watches a video or surfs the internet. We examine the opportunities that can arise from designing a cross- layer adaptive bitrate streaming algorithm. Can a cross-layer algorithm improve the user experience, while not overwhelming the competing traffic flows? i Contents Contents iii 1 Introduction1 2 Background3 2.1 Adaptive Bitrate Streaming.................... 3 2.2 Streaming Algorithms ....................... 5 2.2.1 Rate-based approaches................... 5 2.2.2 Buffer-based approaches.................. 6 2.2.3 Modern algorithms..................... 8 2.3 Congestion Control and ABR................... 9 2.3.1 QUIC and BBR....................... 9 2.3.2 Minerva ........................... 11 2.4 Data-centric approaches...................... 13 2.4.1 Reinforcement and Q-learning.............. 13 2.4.2 Reinforcement learning and congestion control . 15 3 Streaming DASH segments over QUIC 17 3.1 Server-side ABR Decision ..................... 17 3.2 DASH.js player adaptation..................... 19 3.3 Experimental setup......................... 20 4 Designing cross-layer algorithms 23 4.1 Problem formulation........................ 23 4.2 WorthedAbr: VMAF-aware planning............... 25 4.3 TargetAbr: low-liberty long-term planning........... 30 4.4 AutoTarget: generalizing congestion control .......... 34 4.5 GapAbr: specializing congestion control............. 37 4.6 Design space............................. 42 5 Evaluation 45 iii Contents 5.1 Experimental setup......................... 45 5.2 Competitive Scenarios....................... 47 5.2.1 Heterogeneous Videos................... 47 5.2.2 Homogeneous Videos................... 51 5.2.3 Multiple Flows....................... 54 5.3 Performance on Traces....................... 56 5.4 Self Fairness............................. 59 5.5 Background TCP Traffic...................... 63 5.6 Summary............................... 66 6 Conclusions 67 Bibliography 69 iv Chapter 1 Introduction Video traffic makes up most of today’s internet traffic. Yet, most transport layer protocols are oblivious to the nature of the traffic transmitted through them, while video players hardly consider low-level characteristics of the videos. In a typical setup, videos are divided in chunks and streamed at application-level over HTTP connections. However, the underlying trans- port can not tell if the user watches a video or surfs the internet as the video player delivers segment after segment. This thesis aims to examine the opportunities that can arise from designing cross-layer adaptive bitrate streaming algorithms. Since HTTP runs on top of transport-layer protocols, the performance of video players is directly affected both by network conditions and the imple- mentation of the transport layer. Congestion control adjusts the data trans- mission rate in order to avoid and recover from network congestion. We want to asses whether joint optimization of congestion control and adaptive bitrate streaming can lead to superior viewing experience. We continue prior work[18] which shows that communication between the 2 aforementioned layers is indeed beneficial. We take the concept of inter-layer communication one step further: we provide algorithms that jointly optimize for better user experience. Contributions We introduce a framework for taking on-the fly decisions about the qual- ity of the video segments served directly on the server side(chapter3) by exploiting long polling and the features of QUIC[14]. We propose a series of cross layer algorithms(chapter4), from which GapAbr (section 4.5) stands out by achieving better quality of experience results when running on its own(section 5.3). Furthermore, GapAbr acts fair to- wards other ABR algorithms by taking advantage of the video’s structure 1 1. Introduction and adapting accordingly(section 5.4). In chapter4, we also highlight the most important design decisions, along with implementation details and heuristics for achieving this important outcome. Along with Gap, we ex- plore a series of diverse algorithms ranging from short-sight heuristic plan- ning(section 4.2) to reinforcement learning(section 4.5) and control theory (section 4.5). Finally, we qualify both the efficiency and fairness of our solutions by ex- perimenting over 4 different videos in multiple scenarios(chapter5). We compare our algorithms with 3 different benchmarks(RobustMPC[26], Dy- namic [22] and Minerva[18]). We explore both the performance and fairness of Gap in comparison with our baselines in a series of diverse scenarios: iso- lated performance and fairness(section 5.4), performance on traces(section 5.3), competitive scenarios(section 5.2) and against background TCP traf- fic(section 5.5). 2 Chapter 2 Background 2.1 Adaptive Bitrate Streaming Adaptive video streaming is a technique used in streaming video and au- dio media over networks. The approach partitions videos in segments of fixed or variable sizes and uses HTTP to serve them in order. The segments are present for different bitrates to accommodate the varying network con- ditions. An adaptive bit rate(ABR) algorithm can modify the bitrate of the fetched segments dynamically in order to improve the viewing experience. A buffer of segments is stored on the client side, while past segments of the buffer are being consumed by the user that watches the video. Increasing the bitrate of the video improves the experience of the user, but under varying network conditions choosing a higher bitrate may cause the full consumption of the buffer resulting in the video player being stalled. Hence, the ABR algorithm has to balance the opportunity for downloading higher qualities with the risk of causing unwanted rebuffering events. Fur- thermore, experience[15] suggests that sudden changes in bitrates can de- grade the perceived quality of the frames being rendered during the change. Due to this fact, the ABR algorithm has to account for oscillating between qualities too often. MPEG-DASH is the first adaptive bit-rate HTTP-based streaming solution that became an international standard[4]. Alternative segments are encoded at different bit rates covering aligned short intervals of playback time. While the content is being played by the MPEG-DASH client, a bit rate adapta- tion algorithm is used to automatically select the segment with the highest possible rate that allows for playback without causing stalls or re-buffering events. There are multiple factors that users care about when watching a video[26]. 3 2. Background Using higher bitrates entails better video quality, but at the same time in- creases the risk of rebuffering events: not enough downloaded segments cause video to stall. The total duration these events is denoted as rebuffering time. Similarly, the startup time is the time waited by users at the beginning of a video. Finally, ABR algorithms have to take into consideration quality changes: sudden variations in quality can be noticed by users, degrading the experience. In the literature, all the factors mentioned above are combined into diverse Quality of Experience(QoE) metrics[26, 20, 18]. Buffer dynamics are oftentimes formulated under a continuous space and can be succinctly described by the following equations from paper [26]: dk(Rk) tk+1 = tk + + Dtk (2.1) Ck 1 Z tk+1−Dtk Ck = Ct dt (2.2) tk+1 − tk − Dtk tk dk(Rk) Bk+1 = max max Bk − , 0 + L − Dtk, 0 (2.3) Ck where Bk 2 [0, Bmax] represents the buffer occupancy when starting down- th loading the k segment, tk the time when downloading segment k starts and dk(Rk) is the dimension of the segment k encoded at rate Rk. The times Dtk represent the network latency component of the server-side receiving the current segment request, while Bmax represents an upper buffer limitation oftentimes applied to the client buffer. Equation 2.1 describes the download time evolution of the chunks. The time that passes until downloading chunk k + 1(tk+1) is the time needed to download a segment of size dk(Rk) for the corresponding bitrate Rk on top of the time at which segment k started downloading(tk). The download is done over a network capacity for segment k of Ck, that can be approximated from the individual measurements over the interval of time [tk+1 − Dtk, tk] as per equation 2.2. The buffer occupancy Bk+1(equation 2.1) is described by the gain of L seconds from downloading the current segment and the ( ) current buffer B adjusted by the download time of the segment – dk Rk . k Ck The value Dtk describes the timeout between downloading 2 segments, that is half an RTT if the ABR is run on the client-side. Quality of Experience(QoE) is often dependent on various components such as the video quality, rebuffer time, startup time and quality changes[26]. Measuring the QoE is not a straight-forward job as the mapping between the perceived quality of a segment and is not an obvious one, but rather stands to be related to the particular video being watched. Netflix’s approach of measuring the quality of a segment is the Multi-method Assessment Fusion, or VMAF for short. As stated in [15], VMAF is a video quality metric that 4 2.2. Streaming Algorithms combines human vision modeling with machine learning. VMAF captures differences between codecs, as well as scaling artifacts, in a way that’s better correlated with perceptual quality.
Recommended publications
  • Challenges in Relaying Video Back to Mission Control
    Challenges in Relaying Video Back to Mission Control CONTENTS 1 Introduction ..................................................................................................................................................................................... 2 2 Encoding ........................................................................................................................................................................................... 3 3 Time is of the Essence .............................................................................................................................................................. 3 4 The Reality ....................................................................................................................................................................................... 5 5 The Tradeoff .................................................................................................................................................................................... 5 6 Alleviations and Implementation ......................................................................................................................................... 6 7 Conclusion ....................................................................................................................................................................................... 7 White Paper Revision 2 July 2020 By Christopher Fadeley Using a customizable hardware-accelerated encoder is essential to delivering the high
    [Show full text]
  • Hikvision H.264+ Encoding Technology
    WHITE PAPER Hikvision H.264+ Encoding Technology Encoding Improvement / Higher Transmission / Efficiency Storage Savings 2 Contents 1. Introduction .............................................................................................. 3 2. Background ............................................................................................... 3 3. Key Technologies .................................................................................... 4 3.1 Predictive Encoding ........................................................................ 4 3.2 Noise Suppression.......................................................................... 8 3.3 Long-Term Bitrate Control........................................................... 9 4. Applications ............................................................................................ 11 5. Conclusion............................................................................................... 11 Hikvision H.264+ Encoding Technology 3 1. INTRODUCTION As the global market leader in video surveillance products, Hikvision Digital Technology Co., Ltd., continues to strive for enhancement of its products through application of the latest in technology. H.264+ Advanced Video Coding (AVC) optimizes compression beyond the current H.264 standard. Through the combination of intelligent analysis technology with predictive encoding, noise suppression, and long-term bitrate control, Hikvision is meeting the demand for higher resolution at reduced bandwidths. Our customers will benefit
    [Show full text]
  • A Packetization and Variable Bitrate Interframe Compression Scheme for Vector Quantizer-Based Distributed Speech Recognition
    A Packetization and Variable Bitrate Interframe Compression Scheme For Vector Quantizer-Based Distributed Speech Recognition Bengt J. Borgstrom¨ and Abeer Alwan Department of Electrical Engineering, University of California, Los Angeles [email protected], [email protected] Abstract For example, the ETSI standard [2] packetization scheme sim- ply concatenates 2 adjacent 44-bit source coded frames, and We propose a novel packetization and variable bitrate com- transmits the resulting signal along with header information. pression scheme for DSR source coding, based on the Group The proposed algorithm performs further compression using of Pictures concept from video coding. The proposed algorithm interframe correlation. The high time correlation present in simultaneously packetizes and further compresses source coded speech has previously been studied for speech coding [3] and features using the high interframe correlation of speech, and is DSR coding [4]. compatible with a variety of VQ-based DSR source coders. The The proposed algorithm allows for lossless compression of algorithm approximates vector quantizers as Markov Chains, the quantized speech features. However, the algorithm is also and empirically trains the corresponding probability parame- robust to various degrees of lossy compression, either through ters. Feature frames are then compressed as I-frames, P-frames, VQ pruning or frame puncturing. VQ pruning refers to ex- or B-frames, using Huffman tables. The proposed scheme can clusion of low probability VQ codebook labels prior to Huff- perform lossless compression, but is also robust to lossy com- man coding, and thus excludes longer Huffman codewords, pression through VQ pruning or frame puncturing. To illustrate and frame puncturing refers to the non-transmission of certain its effectiveness, we applied the proposed algorithm to the ETSI frames, which drastically reduces the final bitrate.
    [Show full text]
  • AVC to the Max: How to Configure Encoder
    Contents Company overview …. ………………………………………………………………… 3 Introduction…………………………………………………………………………… 4 What is AVC….………………………………………………………………………… 6 Making sense of profiles, levels, and bitrate………………………………………... 7 Group of pictures and its structure..………………………………………………… 11 Macroblocks: partitioning and prediction modes….………………………………. 14 Eliminating spatial redundancy……………………………………………………… 15 Eliminating temporal redundancy……...……………………………………………. 17 Adaptive quantization……...………………………………………………………… 24 Deblocking filtering….….…………………………………………………………….. 26 Entropy encoding…………………………………….……………………………….. 2 8 Conclusion…………………………………………………………………………….. 29 Contact details..………………………………………………………………………. 30 2 www.elecard.com Company overview Elecard company, founded in 1988, is a leading provider of software products for encoding, decoding, processing, monitoring and analysis of video and audio data in 9700 companies various formats. Elecard is a vendor of professional software products and software development kits (SDKs); products for in - depth high - quality analysis and monitoring of the media content; countries 1 50 solutions for IPTV and OTT projects, digital TV broadcasting and video streaming; transcoding servers. Elecard is based in the United States, Russia, and China with 20M users headquarters located in Tomsk, Russia. Elecard products are highly appreciated and widely used by the leaders of IT industry such as Intel, Cisco, Netflix, Huawei, Blackmagic Design, etc. For more information, please visit www.elecard.com. 3 www.elecard.com Introduction Video compression is the key step in video processing. Compression allows broadcasters and premium TV providers to deliver their content to their audience. Many video compression standards currently exist in TV broadcasting. Each standard has different properties, some of which are better suited to traditional live TV while others are more suited to video on demand (VoD). Two basic standards can be identified in the history of video compression: • MPEG-2, a legacy codec used for SD video and early digital broadcasting.
    [Show full text]
  • Adaptive Bitrate Streaming Over Cellular Networks: Rate Adaptation and Data Savings Strategies
    Adaptive Bitrate Streaming Over Cellular Networks: Rate Adaptation and Data Savings Strategies Yanyuan Qin, Ph.D. University of Connecticut, 2021 ABSTRACT Adaptive bitrate streaming (ABR) has become the de facto technique for video streaming over the Internet. Despite a flurry of techniques, achieving high quality ABR streaming over cellular networks remains a tremendous challenge. First, the design of an ABR scheme needs to balance conflicting Quality of Experience (QoE) metrics such as video quality, quality changes, stalls and startup performance, which is even harder under highly dynamic bandwidth in cellular network. Second, streaming providers have been moving towards using Variable Bitrate (VBR) encodings for the video content, which introduces new challenges for ABR streaming, whose nature and implications are little understood. Third, mobile video streaming consumes a lot of data. Although many video and network providers currently offer data saving options, the existing practices are suboptimal in QoE and resource usage. Last, when the audio and video arXiv:2104.01104v2 [cs.NI] 14 May 2021 tracks are stored separately, video and audio rate adaptation needs to be dynamically coordinated to achieve good overall streaming experience, which presents interesting challenges while, somewhat surprisingly, has received little attention by the research community. In this dissertation, we tackle each of the above four challenges. Firstly, we design a framework called PIA (PID-control based ABR streaming) that strategically leverages PID control concepts and novel approaches to account for the various requirements of ABR streaming. The evaluation results demonstrate that PIA outperforms state-of-the-art schemes in providing high average bitrate with signif- icantly lower bitrate changes and stalls, while incurring very small runtime overhead.
    [Show full text]
  • Comments on `Area and Power Efficient DCT Architecture for Image
    Cintra and Bayer EURASIP Journal on Advances in Signal EURASIP Journal on Advances Processing (2017) 2017:50 DOI 10.1186/s13634-017-0486-8 in Signal Processing RESEARCH Open Access Comments on ‘Area and power efficient DCT architecture for image compression’ by Dhandapani and Ramachandran Renato J. Cintra1* and Fábio M. Bayer2 Abstract In [Dhandapani and Ramachandran, “Area and power efficient DCT architecture for image compression”, EURASIP Journal on Advances in Signal Processing 2014, 2014:180] the authors claim to have introduced an approximation for the discrete cosine transform capable of outperforming several well-known approximations in literature in terms of additive complexity. We could not verify the above results and we offer corrections for their work. Keywords: DCT approximations, Low-complexity transforms 1 Introduction 2 Criticisms In a recent paper [1], a low-complexity transformation 2.1 Inverse transformation was introduced, which is claimed to be a good approxima- The authors of [1] claim that inverse transformation T−1 tion to the discrete cosine transform (DCT). We wish to is given by evaluate this claim. ⎡ ⎤ The introduced transformation is given by the following 10001000 ⎢ ⎥ matrix: ⎢ −1100−11 0 0⎥ ⎢ ⎥ ⎢ 00100010⎥ ⎡ ⎤ ⎢ ⎥ 1 ⎢ 00−11 0 0−11⎥ 10000001 · ⎢ ⎥ . ⎢ ⎥ 2 ⎢ 00−11 0 0 1−1 ⎥ ⎢ 11000011⎥ ⎢ ⎥ ⎢ ⎥ ⎢ 001000−10⎥ ⎢ 00100100⎥ ⎣ ⎦ ⎢ ⎥ −11001−10 0 ⎢ 00111100⎥ T = ⎢ ⎥ . 1000−10 0 0 ⎢ 00 1 1−1 −10 0⎥ ⎢ ⎥ ⎢ 00 1 0 0−10 0⎥ ⎣ ⎦ However, simple computation reveal that this is not 110000−1 −1 accurate, being the actual inverse given by: 1000000−1 ⎡ ⎤ 10000001 ⎢ ⎥ We aim at analyzing the above matrix and showing that ⎢ −1100001−1 ⎥ ⎢ ⎥ it does not consist of a meaningful approximation for ⎢ 00100100⎥ ⎢ ⎥ the8-pointDCT.Inthefollowing,weadoptedthesame − 1 ⎢ 00−11 1−10 0⎥ T 1 = · ⎢ ⎥ .
    [Show full text]
  • Neural Rate Control for Video Encoding Using Imitation Learning
    Neural Rate Control for Video Encoding using Imitation Learning Hongzi Mao†? Chenjie Gu† Miaosen Wang† Angie Chen† Nevena Lazic† Nir Levine† Derek Pang* Rene Claus* Marisabel Hechtman* Ching-Han Chiang* Cheng Chen* Jingning Han* †DeepMind *Google ?MIT CSAIL Abstract module manages the distribution of available bandwidth, usually constrained by the network condition. It directly In modern video encoders, rate control is a critical com- determines how many bits to spend to encode each video ponent and has been heavily engineered. It decides how frame, by assigning a quantization parameter (QP). The many bits to spend to encode each frame, in order to opti- goal of rate control is to maximize the encoding efficiency mize the rate-distortion trade-off over all video frames. This (often measured by the Bjontegaard delta rate) as well as is a challenging constrained planning problem because of maintaining the bitrate under a user-specified target bitrate. the complex dependency among decisions for different video Rate control is a constrained planning problem and can frames and the bitrate constraint defined at the end of the be formulated as a Partially Observable Markov Decision episode. Process. For a particular frame, the QP assignment decision We formulate the rate control problem as a Partially depends on (1) the frame’s spatial and temporal complexity, Observable Markov Decision Process (POMDP), and ap- (2) the previously encoded frames this frame refers to, and ply imitation learning to learn a neural rate control pol- (3) the complexity of future frames. In addition, the bitrate icy. We demonstrate that by learning from optimal video constraint imposes an episodic constraint on the planning encoding trajectories obtained through evolution strategies, problem.
    [Show full text]
  • Forensic Reconstruction of Fragmented Variable Bitrate MP3 Files
    CORE Metadata, citation and similar papers at core.ac.uk Provided by ScholarWorks @ The University of New Orleans University of New Orleans ScholarWorks@UNO University of New Orleans Theses and Dissertations Dissertations and Theses 12-17-2010 Forensic Reconstruction of Fragmented Variable Bitrate MP3 files Abhilash Sajja University of New Orleans Follow this and additional works at: https://scholarworks.uno.edu/td Recommended Citation Sajja, Abhilash, "Forensic Reconstruction of Fragmented Variable Bitrate MP3 files" (2010). University of New Orleans Theses and Dissertations. 1258. https://scholarworks.uno.edu/td/1258 This Thesis is protected by copyright and/or related rights. It has been brought to you by ScholarWorks@UNO with permission from the rights-holder(s). You are free to use this Thesis in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights- holder(s) directly, unless additional rights are indicated by a Creative Commons license in the record and/or on the work itself. This Thesis has been accepted for inclusion in University of New Orleans Theses and Dissertations by an authorized administrator of ScholarWorks@UNO. For more information, please contact [email protected]. Forensic Reconstruction of Fragmented Variable Bitrate MP3 files A Thesis Submitted to the Graduate Faculty of the University of New Orleans In partial fulfillment of the requirements for the degree of Master of Science in Computer Science by Abhilash Sajja December, 2010 Acknowledgements First I would like to thank my parents for their support and encouragement. Special thanks to my advisor Dr.Vassil Roussev for his guidance and feedback without which this thesis would not have been possible.
    [Show full text]
  • A/330, "Link Layer Protocol"
    ATSC A/330:2016 Link-Layer Protocol 19 September 2016 ATSC Standard: Link-Layer Protocol (A/330) Doc. A/330:2016 19 September 2016 Advanced Television Systems Committee 1776 K Street, N.W. Washington, D.C. 20006 202-872-9160 i ATSC A/330:2016 Link-Layer Protocol 19 September 2016 The Advanced Television Systems Committee, Inc., is an international, non-profit organization developing voluntary standards and recommended practices for digital television. ATSC member organizations represent the broadcast, broadcast equipment, motion picture, consumer electronics, computer, cable, satellite, and semiconductor industries. ATSC also develops digital television implementation strategies and supports educational activities on ATSC standards. ATSC was formed in 1983 by the member organizations of the Joint Committee on InterSociety Coordination (JCIC): the Electronic Industries Association (EIA), the Institute of Electrical and Electronic Engineers (IEEE), the National Association of Broadcasters (NAB), the National Cable Telecommunications Association (NCTA), and the Society of Motion Picture and Television Engineers (SMPTE). For more information visit www.atsc.org. Note: The user's attention is called to the possibility that compliance with this standard may require use of an invention covered by patent rights. By publication of this standard, no position is taken with respect to the validity of this claim or of any patent rights in connection therewith. One or more patent holders have, however, filed a statement regarding the terms on which such patent holder(s) may be willing to grant a license under these rights to individuals or entities desiring to obtain such a license. Details may be obtained from the ATSC Secretary and the patent holder.
    [Show full text]
  • A Psychoacoustic-Based Multiple Audio Object Coding Approach Via Intra- Object Sparsity
    University of Wollongong Research Online Faculty of Engineering and Information Faculty of Engineering and Information Sciences - Papers: Part B Sciences 2017 A psychoacoustic-based multiple audio object coding approach via intra- object sparsity Maoshen Jia Beijing University of Technology Jiaming Zhang Beijing University of Technology Changchun Bao Beijing University of Technology, [email protected] Xiguang Zheng University of Wollongong, Dolby Laboratories, [email protected] Follow this and additional works at: https://ro.uow.edu.au/eispapers1 Part of the Engineering Commons, and the Science and Technology Studies Commons Recommended Citation Jia, Maoshen; Zhang, Jiaming; Bao, Changchun; and Zheng, Xiguang, "A psychoacoustic-based multiple audio object coding approach via intra-object sparsity" (2017). Faculty of Engineering and Information Sciences - Papers: Part B. 882. https://ro.uow.edu.au/eispapers1/882 Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected] A psychoacoustic-based multiple audio object coding approach via intra-object sparsity Abstract Rendering spatial sound scenes via audio objects has become popular in recent years, since it can provide more flexibility for different auditory scenarios, such as 3D movies, spatial audio communication and virtual classrooms. To facilitate high-quality bitrate-efficient distribution for spatial audio objects, an encoding scheme based on intra-object sparsity (approximate k-sparsity of the audio object itself) is proposed in this paper. The statistical analysis is presented to validate the notion that the audio object has a stronger sparseness in the Modified Discrete Cosine Transform (MDCT) domain than in the Short Time Fourier Transform (STFT) domain.
    [Show full text]
  • NETWORK VIDEO ENCODER User Manual
    NETWORK VIDEO ENCODER User Manual SPE-110 Network Video Encoder User Manual Copyright ©2018 Hanwha Techwin Co., Ltd. All rights reserved. Trademark Each of trademarks herein is registered. The name of this product and other trademarks mentioned in this manual are the registered trademark of their respective company. Restriction Copyright of this document is reserved. Under no circumstances, this document shall be reproduced, distributed or changed, partially or wholly, without formal authorization. Disclaimer Hanwha Techwin makes the best to verify the integrity and correctness of the contents in this document, but no formal guarantee shall be provided. Use of this document and the subsequent results shall be entirely on the user’s own responsibility. Hanwha Techwin reserves the right to change the contents of this document without prior notice. Design and specifications are subject to change without prior notice. The initial administrator ID is “admin” and the password should be set when logging in for the first time. Please change your password every three months to safely protect personal information and to prevent the damage of the information theft. Please, take note that it’s a user’s responsibility for the security and any other problems caused by mismanaging a password. overview IMPORTANT SAFETY INSTRUCTIONS WARNING TO REDUCE THE RISK OF FIRE OR ELECTRIC SHOCK, DO NOT EXPOSE THIS PRODUCT 1. Read these instructions. TO RAIN OR MOISTURE. DO NOT INSERT ANY METALLIC OBJECT THROUGH THE 2. Keep these instructions. VENTILATION GRILLS OR OTHER OPENNINGS ON THE EQUIPMENT. ● Heed all warnings. ● 3. Apparatus shall not be exposed to dripping or splashing and that no objects filled with liquids, OVERVIEW 4.
    [Show full text]
  • Compressing Variational Posteriors
    Compressing Variational Posteriors Yibo Yang∗ Robert Bamler∗ Stephan Mandt UC Irvine UC Irvine UC Irvine [email protected] [email protected] [email protected] Abstract Deep Bayesian latent variable models have enabled new approaches to both model and data compression. Here, we propose a new algorithm for compressing latent representations in deep probabilistic models, such as VAEs, in post-processing. Our algorithm generalizes arithmetic coding to the continuous domain, using adaptive discretization accuracy that exploits posterior uncertainty. Our approach separates model design and training from the compression task, and thus allows for various rate-distortion trade-offs with a single trained model, eliminating the need to train multiple models for different bit rates. We obtain promising experimental results on compressing Bayesian neural word embeddings, and outperform JPEG on image compression over a wide range of bit rates using only a single standard VAE. 1 Introduction and Related Work One natural application of deep generative modeling is data compression. Here we focus on lossy compression of continuous data, which entails quantization, i.e., discretizing arbitrary real numbers to a countable number of representative values, resulting in the classic trade-off between rate (entropy of the discretized values) and distortion (error from discretization). State-of-the-art deep learning based approaches to lossy data compression [1–5] train a machine learning model to perform lossy compression by optimizing a rate-distortion objective, substituting the quantization step with a differentiable approximation during training. Consequently, a new model has to be trained for any desired combination of rate and distortion. To support variable-bitrate compression, one has to train and store several deep learning models for different rates, resulting in possibly very large codec size.
    [Show full text]