search

HongminWu / BP-AR-HMM

This repository is used to discover and model dynamicl behaviors which are shared among several related time series.
8 stars 3 forks source link

copyright of this repo #1

Open huiwenzhang opened 6 years ago

huiwenzhang commented 6 years ago

Hi, I want to know if this repo comes from E. B. Fox and modified by Scott Niekum? Because I searched the website and found no information about this package in Scott Niekum's personal homepage or his paper. Actually, I want to reproduce Scott Niekum's experiment in one of his paper, so I need to confirm he used this package in his experiment. Any feedback is welcomed. Thank you.

HongminWu commented 6 years ago

Yes, this repo was initially shared by Scott Niekum for his IJRR2015 paper-"Learning Grounded Finite-State Representations from Unstructured Demonstration". You can find the code in Scott Niekum's personal homepage by

  1. Access Home page: https://www.cs.utexas.edu/~sniekum/index.php
  2. Click PUBLICATIONS
  3. Scroll down until the publications in 2015, and then you can find the code attached with the corresponding paper.
  4. In the code page https://www.cs.utexas.edu/~sniekum/code.php, you should find the downloading link under the Other code section, named,
    • Matlab code* for automatic segmentation of demonstrations using the BP-AR-HMM.

Hopefully, this can help you.

On Thu, 4 Oct 2018 at 00:17, Alvin Zhang notifications@github.com wrote:

Hi, I want to know if this repo comes from E. B. Fox and modified by Scott Niekum? Because I searched the website and found no information about this package in Scott Niekum's personal homepage or his paper. Actually, I want to reproduce Scott Niekum's experiment in one of his paper, so I need to confirm he used this package in his experiment. Any feedback is welcomed. Thank you.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/HongminWu/BP-AR-HMM/issues/1, or mute the thread https://github.com/notifications/unsubscribe-auth/ALLZeV4nStmYMooSmuaWn3Lu8ewDxtrIks5uhOMSgaJpZM4XGYUu .

-- ​ Hongmin Wu, Ph.D. Candidate Biomimetic and Intelligent Robotics Lab. (BIRL) School of Electromechanical Engineering Guangdong University of Technology Guangzhou, Guangdong, P.R.China 510006 WeChat: Homing20 Facebook: Hongmin Wu Tel:86+18819498204 E-mail:hongminwu0120@gmail.com hongminwu0120@gmail.com URL: https://hongminwu.github.io https://hongminwu.github.io/ http://www.juanrojas.net/research/​

huiwenzhang commented 6 years ago

Yes, this repo was initially shared by Scott Niekum for his IJRR2015 paper-"Learning Grounded Finite-State Representations from Unstructured Demonstration". You can find the code in Scott Niekum's personal homepage by 1. Access Home page: https://www.cs.utexas.edu/~sniekum/index.php 2. Click PUBLICATIONS 3. Scroll down until the publications in 2015, and then you can find the code attached with the corresponding paper. 3. In the code page https://www.cs.utexas.edu/~sniekum/code.php, you should find the downloading link under the Other code section, named, Matlab code for automatic segmentation of demonstrations using the BP-AR-HMM. Hopefully, this can help you. … On Thu, 4 Oct 2018 at 00:17, Alvin Zhang @.**> wrote: Hi, I want to know if this repo comes from E. B. Fox and modified by Scott Niekum? Because I searched the website and found no information about this package in Scott Niekum's personal homepage or his paper. Actually, I want to reproduce Scott Niekum's experiment in one of his paper, so I need to confirm he used this package in his experiment. Any feedback is welcomed. Thank you. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1>, or mute the thread https://github.com/notifications/unsubscribe-auth/ALLZeV4nStmYMooSmuaWn3Lu8ewDxtrIks5uhOMSgaJpZM4XGYUu . -- ​ Hongmin Wu, Ph.D. Candidate Biomimetic and Intelligent Robotics Lab. (BIRL) School of Electromechanical Engineering Guangdong University of Technology Guangzhou, Guangdong, P.R.China 510006 WeChat: Homing20 Facebook: Hongmin Wu Tel:86+18819498204 E-mail:hongminwu0120@gmail.com hongminwu0120@gmail.com URL: https://hongminwu.github.io https://hongminwu.github.io/ <http://www.juanrojas.net/research/>​*

Thanks, it is really helpful. By the way, I noticed that there is a demo folder in your repo, which I can't find from Scott's version. Did you update the package ?

HongminWu commented 6 years ago

No, Actually, I didn't update anything. I just applied it to my case, which for segmenting a robot assembly task.

By the way, what's your purpose for using the BP-AR-HMM ?

On Thu, 4 Oct 2018 at 09:46, Alvin Zhang notifications@github.com wrote:

Yes, this repo was initially shared by Scott Niekum for his IJRR2015 paper-"Learning Grounded Finite-State Representations from Unstructured Demonstration". You can find the code in Scott Niekum's personal homepage by 1. Access Home page: https://www.cs.utexas.edu/~sniekum/index.php 2. Click PUBLICATIONS 3. Scroll down until the publications in 2015, and then you can find the code attached with the corresponding paper. 3. In the code page https://www.cs.utexas.edu/~sniekum/code.php, you should find the downloading link under the Other code section, named, Matlab code for automatic segmentation of demonstrations using the BP-AR-HMM. Hopefully, this can help you. … <#m1239136080824590440> On Thu, 4 Oct 2018 at 00:17, Alvin Zhang @.**> wrote: Hi, I want to know if this repo comes from E. B. Fox and modified by Scott Niekum? Because I searched the website and found no information about this package in Scott Niekum's personal homepage or his paper. Actually, I want to reproduce Scott Niekum's experiment in one of his paper, so I need to confirm he used this package in his experiment. Any feedback is welcomed. Thank you. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1 https://github.com/HongminWu/BP-AR-HMM/issues/1>, or mute the thread https://github.com/notifications/unsubscribe-auth/ALLZeV4nStmYMooSmuaWn3Lu8ewDxtrIks5uhOMSgaJpZM4XGYUu . -- ​ Hongmin Wu, Ph.D. Candidate Biomimetic and Intelligent Robotics Lab. (BIRL) School of Electromechanical Engineering Guangdong University of Technology Guangzhou, Guangdong, P.R.China 510006 WeChat: Homing20 Facebook: Hongmin Wu Tel:86+18819498204 E-mail:hongminwu0120@gmail.com hongminwu0120@gmail.com hongminwu0120@gmail.com hongminwu0120@gmail.com URL: https://hongminwu.github.io https://hongminwu.github.io/ < http://www.juanrojas.net/research/>​*

Thanks, it is really helpful. By the way, I noticed that there is a demo folder in your repo, which I can't find from Scott's version. Did you update the package ?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/HongminWu/BP-AR-HMM/issues/1#issuecomment-426858441, or mute the thread https://github.com/notifications/unsubscribe-auth/ALLZedG_yXVIiKQQ6nowBuGO9IRA3Nu2ks5uhWhfgaJpZM4XGYUu .

-- ​ Hongmin Wu, Ph.D. Candidate Biomimetic and Intelligent Robotics Lab. (BIRL) School of Electromechanical Engineering Guangdong University of Technology Guangzhou, Guangdong, P.R.China 510006 WeChat: Homing20 Facebook: Hongmin Wu Tel:86+18819498204 E-mail:hongminwu0120@gmail.com hongminwu0120@gmail.com URL: https://hongminwu.github.io https://hongminwu.github.io/ http://www.juanrojas.net/research/​

huiwenzhang commented 6 years ago

Well, probably the same purpose. I use it to segment a sequential task in the context of imitation learning. The difference is the BP-AR-HMM serves as a contrast experiment, which will be compared with another algorithm. Cause you know BP-AR-HMM is kind of the state of art performance for nonparametric segmentation, at least theoretically. How about your results? Is it good for your assembly task. I guess these methods maybe task related and more critically, it is hard to say what is a good segment. Finally, if you are interested in my work, we can contact offline.