Collaborative Intent Prediction with Real-Time Contextual Data

2017 ◽  
Vol 35 (4) ◽  
pp. 1-33 ◽  
Author(s):  
Yu Sun ◽  
Nicholas Jing Yuan ◽  
Xing Xie ◽  
Kieran McDonald ◽  
Rui Zhang
Keyword(s):  
Real Time ◽  
Contextual Data ◽  
Intent Prediction

EdgeDIPN

2020 ◽  
Vol 14 (3) ◽  
pp. 320-328
Author(s):  
Long Guo ◽  
Lifeng Hua ◽  
Rongfei Jia ◽  
Fei Fang ◽  
Binqiang Zhao ◽  
...  
Keyword(s):  
Real Time ◽  
Large Scale ◽  
Attention Mechanism ◽  
Data Sources ◽  
User Intent ◽  
Multiple User ◽  
Shopping Experience ◽  
Data Source ◽  
Intent Prediction ◽  
The Right

With the rapid growth of e-commerce in recent years, e-commerce platforms are becoming a primary place for people to find, compare and ultimately purchase products. To improve online shopping experience for consumers and increase sales for sellers, it is important to understand user intent accurately and be notified of its change timely. In this way, the right information could be offered to the right person at the right time. To achieve this goal, we propose a unified deep intent prediction network, named EdgeDIPN, which is deployed at the edge, i.e., mobile device, and able to monitor multiple user intent with different granularity simultaneously in real-time. We propose to train EdgeDIPN with multi-task learning, by which EdgeDIPN can share representations between different tasks for better performance and saving edge resources in the meantime. In particular, we propose a novel task-specific attention mechanism which enables different tasks to pick out the most relevant features from different data sources. To extract the shared representations more effectively, we utilize two kinds of attention mechanisms, where the multi-level attention mechanism tries to identify the important actions within each data source and the inter-view attention mechanism learns the interactions between different data sources. In the experiments conducted on a large-scale industrial dataset, EdgeDIPN significantly outperforms the baseline solutions. Moreover, EdgeDIPN has been deployed in the operational system of Alibaba. Online A/B testing results in several business scenarios reveal the potential of monitoring user intent in real-time. To the best of our knowledge, EdgeDIPN is the first full-fledged real-time user intent understanding center deployed at the edge and serving hundreds of millions of users in a large-scale e-commerce platform.

Spatio-temporal DenseNet for real-time intent prediction of pedestrians in urban traffic environments

2020 ◽  
Vol 386 ◽  
pp. 317-324
Author(s):  
Khaled Saleh ◽  
Mohammed Hossny ◽  
Saeid Nahavandi
Keyword(s):  
Real Time ◽  
Urban Traffic ◽  
Spatio Temporal ◽  
Intent Prediction

scholarly journals Myoelectric digit action decoding with multi-output, multi-class classification: an offline analysis

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Agamemnon Krasoulis ◽  
Kianoush Nazarpour
Keyword(s):  
Real Time ◽  
Degrees Of Freedom ◽  
Real Time Control ◽  
Artificial Joints ◽  
Time Control ◽  
Offline Analysis ◽  
Intent Prediction ◽  
Multi Class Classification ◽  
Prosthetic Finger ◽  
Finger Control

Abstract The ultimate goal of machine learning-based myoelectric control is simultaneous and independent control of multiple degrees of freedom (DOFs), including wrist and digit artificial joints. For prosthetic finger control, regression-based methods are typically used to reconstruct position/velocity trajectories from surface electromyogram (EMG) signals. Unfortunately, such methods have thus far met with limited success. In this work, we propose action decoding, a paradigm-shifting approach for independent, multi-digit movement intent prediction based on multi-output, multi-class classification. At each moment in time, our algorithm decodes movement intent for each available DOF into one of three classes: open, close, or stall (i.e., no movement). Despite using a classifier as the decoder, arbitrary hand postures are possible with our approach. We analyse a public dataset previously recorded and published by us, comprising measurements from 10 able-bodied and two transradial amputee participants. We demonstrate the feasibility of using our proposed action decoding paradigm to predict movement action for all five digits as well as rotation of the thumb. We perform a systematic offline analysis by investigating the effect of various algorithmic parameters on decoding performance, such as feature selection and choice of classification algorithm and multi-output strategy. The outcomes of the offline analysis presented in this study will be used to inform the real-time implementation of our algorithm. In the future, we will further evaluate its efficacy with real-time control experiments involving upper-limb amputees.

Some Recent Results in Quiescent Prominence Spectrometry

1979 ◽  
Vol 44 ◽  
pp. 41-47
Author(s):  
Donald A. Landman
Keyword(s):  
Real Time ◽  
Computer System ◽  
Spectral Response ◽  
Absolute Intensity ◽  
Solar Observatory ◽  
Target Size ◽  
Small Target ◽  
Signal Averaging ◽  
Quiescent Prominence

This paper describes some recent results of our quiescent prominence spectrometry program at the Mees Solar Observatory on Haleakala. The observations were made with the 25 cm coronagraph/coudé spectrograph system using a silicon vidicon detector. This detector consists of 500 contiguous channels covering approximately 6 or 80 Å, depending on the grating used. The instrument is interfaced to the Observatory’s PDP 11/45 computer system, and has the important advantages of wide spectral response, linearity and signal-averaging with real-time display. Its principal drawback is the relatively small target size. For the present work, the aperture was about 3″ × 5″. Absolute intensity calibrations were made by measuring quiet regions near sun center.

Video real-time imaging of artificial membranes during freeze-drying by cryo-electron microscopy

1988 ◽  
Vol 46 ◽  
pp. 16-17
Author(s):  
Alan S. Rudolph ◽  
Ronald R. Price
Keyword(s):  
Real Time ◽  
Self Assembly ◽  
Freeze Drying ◽  
Video Capture ◽  
Drying Process ◽  
Artificial Membranes ◽  
The Past ◽  
Cryo Electron Microscopy ◽  
Spherical Structures ◽  
Capture Techniques

We have employed cryoelectron microscopy to visualize events that occur during the freeze-drying of artificial membranes by employing real time video capture techniques. Artificial membranes or liposomes which are spherical structures within internal aqueous space are stabilized by water which provides the driving force for spontaneous self-assembly of these structures. Previous assays of damage to these structures which are induced by freeze drying reveal that the two principal deleterious events that occur are 1) fusion of liposomes and 2) leakage of contents trapped within the liposome [1]. In the past the only way to access these events was to examine the liposomes following the dehydration event. This technique allows the event to be monitored in real time as the liposomes destabilize and as water is sublimed at cryo temperatures in the vacuum of the microscope. The method by which liposomes are compromised by freeze-drying are largely unknown. This technique has shown that cryo-protectants such as glycerol and carbohydrates are able to maintain liposomal structure throughout the drying process.

An Interactive Minicomputer Program for the Indexing and Simulation of Electron Diffraction Patterns

1976 ◽  
Vol 34 ◽  
pp. 542-543
Author(s):  
R.P. Goehner ◽  
W.T. Hatfield ◽  
Prakash Rao
Keyword(s):  
Electron Diffraction ◽  
Real Time ◽  
Pattern Analysis ◽  
Flow Diagram ◽  
Hard Copy ◽  
Time Analysis ◽  
Real Time Analysis ◽  
Transmission Electron ◽  
One Step ◽  
Diffraction Patterns

Computer programs are now available in various laboratories for the indexing and simulation of transmission electron diffraction patterns. Although these programs address themselves to the solution of various aspects of the indexing and simulation process, the ultimate goal is to perform real time diffraction pattern analysis directly off of the imaging screen of the transmission electron microscope. The program to be described in this paper represents one step prior to real time analysis. It involves the combination of two programs, described in an earlier paper(l), into a single program for use on an interactive basis with a minicomputer. In our case, the minicomputer is an INTERDATA 70 equipped with a Tektronix 4010-1 graphical display terminal and hard copy unit.A simplified flow diagram of the combined program, written in Fortran IV, is shown in Figure 1. It consists of two programs INDEX and TEDP which index and simulate electron diffraction patterns respectively. The user has the option of choosing either the indexing or simulating aspects of the combined program.

Microstructural investigation of surface roughness in MBE-grown AlAs

1995 ◽  
Vol 53 ◽  
pp. 454-455
Author(s):  
R. Rajesh ◽  
R. Droopad ◽  
C. H. Kuo ◽  
R. W. Carpenter ◽  
G. N. Maracas
Keyword(s):  
Thin Film ◽  
Real Time ◽  
Growth Control ◽  
Native Oxide ◽  
Effusion Cell ◽  
Surface Oxide Layer ◽  
Microstructural Investigation ◽  
Mbe Growth ◽  
Film Substrate ◽  
And Control

Knowledge of material pseudodielectric functions at MBE growth temperatures is essential for achieving in-situ, real time growth control. This allows us to accurately monitor and control thicknesses of the layers during growth. Undesired effusion cell temperature fluctuations during growth can thus be compensated for in real-time by spectroscopic ellipsometry. The accuracy in determining pseudodielectric functions is increased if one does not require applying a structure model to correct for the presence of an unknown surface layer such as a native oxide. Performing these measurements in an MBE reactor on as-grown material gives us this advantage. Thus, a simple three phase model (vacuum/thin film/substrate) can be used to obtain thin film data without uncertainties arising from a surface oxide layer of unknown composition and temperature dependence.In this study, we obtain the pseudodielectric functions of MBE-grown AlAs from growth temperature (650°C) to room temperature (30°C). The profile of the wavelength-dependent function from the ellipsometry data indicated a rough surface after growth of 0.5 μm of AlAs at a substrate temperature of 600°C, which is typical for MBE-growth of GaAs.

Real-time observation of vortex lattices in a superconductor

1993 ◽  
Vol 51 ◽  
pp. 1050-1051
Author(s):  
K. Harada ◽  
T. Matsuda ◽  
J.E. Bonevich ◽  
M. Igarashi ◽  
S. Kondo ◽  
...  
Keyword(s):  
Real Time ◽  
Flux Line ◽  
Electron Holography ◽  
Lorentz Microscopy ◽  
Vortex Lattices ◽  
Flux Lines ◽  
Magnetic Flux Lines ◽  
Time Observation ◽  
Thin Superconductor ◽  
Real Time Observation

Previous observations of magnetic flux-lines (vortex lattices) in superconductors, such as the field distribution of a flux-line, and flux-line dynamics activated by heat and current, have employed the high spatial resolution and magnetic sensitivity of electron holography. And recently, the 2-D static distribution of vortices was also observed by this technique. However, real-time observations of the vortex lattice, in spite of scientific and technological interest, have not been possible due to experimental difficulties. Here, we report the real-time observation of vortex lattices in a thin superconductor, by means of Lorentz microscopy using a 300 kV field emission electron microscope. This technique allows us to observe the dynamic motion of individual vortices and record the events on a VTR system.The experimental arrangement is shown in Fig. 1. A Nb thin film for transmission observation was prepared by chemical etching. The grain size of the film was increased by annealing, and single crystals were observed with a thickness of 50∼90 nm.

Live Confocal Analysis of Fertilization and Early Development

2001 ◽  
Vol 7 (S2) ◽  
pp. 1012-1013
Author(s):  
Uyen Tram ◽  
William Sullivan
Keyword(s):  
Drosophila Melanogaster ◽  
Embryonic Development ◽  
Real Time ◽  
Early Development ◽  
Fluorescent Probes ◽  
Primary Defect ◽  
Fluorescent Analysis ◽  
Dynamic Event ◽  
Enormous Amount ◽  
Fluorescent Studies

Embryonic development is a dynamic event and is best studied in live animals in real time. Much of our knowledge of the early events of embryogenesis, however, comes from immunofluourescent analysis of fixed embryos. While these studies provide an enormous amount of information about the organization of different structures during development, they can give only a static glimpse of a very dynamic event. More recently real-time fluorescent studies of living embryos have become much more routine and have given new insights to how different structures and organelles (chromosomes, centrosomes, cytoskeleton, etc.) are coordinately regulated. This is in large part due to the development of commercially available fluorescent probes, GFP technology, and newly developed sensitive fluorescent microscopes. For example, live confocal fluorescent analysis proved essential in determining the primary defect in mutations that disrupt early nuclear divisions in Drosophila melanogaster. For organisms in which GPF transgenics is not available, fluorescent probes that label DNA, microtubules, and actin are available for microinjection.

Sign in / Sign up

Export Citation Format

Share Document