Over the past few months, I have been working with Ferdi on machine learning. We are exploring the capabilities of different machine learning libraries like TensorFlow. In the past, I have worked as a Shop Manager on the lathe; to see more about my contributions with that, go to the Lathe’s project page here.
Week – May 7, 2018
I have got the Barometric and Temperature sensor working on the 9v battery so it can record data and transfer it to the micro sd card without the use of a computer. The code is uploaded to the Arduino and under the assets folder on the WISRD server for backup. I have begun a test to see how long the 9v battery will last. I have started recording at 1:56 pm on May 11. The next steps are cut the wires shorter and solder them to the micro sd card reader and aruino so we dont need the breadboard. We also need to coordinate with Pamela about her sensor and if it will be on this Arduino.
Week- Feb. 19, 2018 – April 9, 2018
Over the past few weeks, I have been working with Ferdi to try to set up an RFID system that allows certain people to use certain machines. We started to work with an old PC, first reinstalling the OS and connecting it the network. With poster presentations coming up we wanted to find something that we could present so we decided to take a break with the RFID and look at sensors. Right now we are seeing what we can do with a temperature and humidity sensor, possibly installing it for the hydroponics team so they can monitor the environment around their plants. We are following the guide here.
Week-Feb. 12, 2018
The NAO support team has gotten back to me and said that it would have to be shipped to for repairs. The full repair could cost upwards of $500. I have forwarded Joe the email and will wait to see how it gets handled. I don’t think my paygrade includes me knowing the WISRD financial details. In the meantime I am working in Coreography, trying to get accustomed to using it.
Week-Feb. 5, 2018
Starting off this week Ferdi and I are trying to do a factory reset on Lucy. We started at the beginning of class but have had no luck in the boot process finishing. It is supposed to take 15-20 minutes so we may have to try again next class. I emailed the support team back and am waiting for a reply.
Week – Jan. 29, 2018
This week I am starting off by trouble shotting the NAO robot. I have learned about how the boot process works and what the LED is telling me about the robot. For instance, right now the eye LEDs are blinking blue, red, green. This means that the robot is booting. The main status LED is fading from off(black) to white. This means that it is booting as well. If the main status LED were blinking blue it would mean that it is upgrading the software and if it were red it would mean the robot is unusable for whatever reason. The serial number for the NAO robot is ALDR1109F00. I submitted a support ticket so I will wait for a reply. After waiting almost 20 more minutes Lucy turned on and stood up. I got her IP address: 192.168.168.70. Her username and password is: nao. On startup Lucy said “Warning 205, data partition may be corrupted. Next step is to connect it to Choreography and give it commands.
Week – Jan. 22, 2018
Starting off this week Ferdi and I are looking into the applications of NAO and I want to look into the specific software that NAO uses. This will allow us to see specifically what limitations there are with the robot. I learned that the software that the robot runs off it NAOqi OS. I remember it from the Youtube videos but will look at it further. I found this documentation for the NAO robot here. I also found the download link to the software here. We have taken lucy out of her box and are running through her boot process. We are learning along the way how the robot operates.
Week – Jan. 15, 2018
Out of school this week – sick.
Week – Jan. 8, 2018
This week Ferdi and I are looking at how we could integrate machine learning with the NAO robot. I learned from Joe that the robot already has some basic things like object recognition and facial recognition. I remember hearing about how previously they used the robot to detect people. Watching the videos on the WISRD youtube channel I learned about how we used it to give out messages copying humanisms. I am interested in how we could integrate its sensors to further copying humanisms.
Week – Nov. 13 – Dec. 4
These past few weeks have been pretty uneventful, but Ferdi and I have found a way to continue our project. We plan to research different machine learning libraries like TensorFlow to see what we can do with them. Our end goal is to have maybe a research paper by the end of the year that shows our research and maybe an example of it working. Our thoughts for this example were to either continue with our current project with twitter emotions or try something different by trying to have the computer program produce its own music.
Week – Nov. 6, 2017
This week was the WISRD poster presentations, everyone in the institute presented what they have been working on for the past few months or longer. Ferdi and I took this chance to explain to some people who were interested, what AI and Machine learning is and how our project works within that. The other class times were spent rethinking our project a little. We want to make sure that before we commit 100% we have a concrete idea, one that will provide value to the institute. Some ideas we had were to look at a piece of text where the author has directly stated what emotion they are portraying. An example is on Facebook, how people can associate an update with a feeling, ie. “feeling happy”. We are still going to research more ways on how to improve.
Week – Oct. 30, 2017
Being a short week with only one class, I did a lot of work on Ferdi and me’s poster for the poster presentation series on Monday, Nov. 5. I also submitted this poster to be printed by Joe. Next week I plan to spend more time rating tweets, trying to break 1000 total within the next two weeks.
Week – Oct. 23, 2017
This week Ferdi and I spent most of our time identifying the emotions in tweets. We got a good deal of tweets done and will hopefully be done by next week. With the tweets we identify, we can train the machine that will hopefully be able to identify the emotion behind tweets its self.
Week – Sep. 11, 2017 – Oct. 16, 2017
These past few weeks Ferdi and I have been working on the twitter scraper. This will allow us to rate tweets based off of what we think the emotion is. Based on our methodology, we will only accept the tweets where we both agree on the emotion for training. This will alleviate confusion on what emotion a tweet is. He has made it so that I can connect to the MySQL database on his computer so now in class I can rate what the tweet it. Looking forward some of my goals are to continue working on the white paper and starting to rate tweets. Hopefully, soon we will have enough data to train the system.
Week – Sep. 4, 2017
This week, Ferdi and I furthered our research on our machine learning project. We started to write our white paper and figure out exactly what we are doing. Recently, we were looking into the classification of emotions and what emotions we wanted to use. We decided to use emotions from Shaver et al. (2001) list.
Week – Aug. 28
This week I spent my time with Ferdi. We were brainstorming ideas for a possible deep learning project. With Ferdi’s coding experience, we wanted to create some kind of program that could take your twitter feed and determine the emotion each tweet is conveying. This would be done through machine learning where we would give a program training data that it would learn from and then produce its own results from what it ‘knows’. In the coming weeks, we plan to figure out and research how it would all work out.
Week-Aug. 25, 2017
This first week of my senior year at WISRD was filled with a lot of orientation for the freshmen. We went over a lot of the expectations/how the Institute model works. I learned how to log on to the server for the first time today, learning how to access the WISRD server from home. I am very excited to get back to working with the lathe and helping out in the shop. I have some things that I think would be interesting to look into for the lathe. I was looking at digital readouts earlier and saw that they could be useful to machining. They allow you to know exactly where your tool bit is in relation to your workpiece. I also want to look into getting a fast swap tool post. This would make it easier to change tool bits while machining.
Week- March 6, 2017 – May 19, 2017
It appears writing this that a big chunk of journal has somehow disappeared or been deleted. I will try to fill in what was written before to ht best of my ability.Over the past few months I have been mainly working on the lathe as I have been the whole year. I have also taken on a few side projects. The other members of the lathe group and I haven’t been asked to manufacture many other pieces but we have had few requests. One of the jobs we were put on was to create inner threads that would accept a 2L pop bottle. We never attempted to do inner threads so we took to both the internet and the old machining books that Joe had. With that information, we were able to complete the piece with working inner threads. We got some new tools and wanted to try them out. One of the new tools was a cut off bar. With this we are able to cut off our work, leaving a square, flat piece. We first tried this with an unknown metal that we think is some sort of alloy. We noticed that using the cutoff bar created a lot of noise. We found that working the lathe at higher speeds made it quieter. When we were using some of the new tools on the lathe, we noticed a sound coming from the motor. The belt holder from the motor came off and the set screw flew off somewhere. We lost the set screw and had to use another one. This time we used medium strength thread locker to ensure that the vibrations from the motor didn’t make the set screw come loose again. On the side I also helped to take a part a old printer. It was to be salvaged for parts, more specifically stepper motors.
Week- February 20, 2017 – March 6, 2017
These past few weeks I have been working on the lathe with Blake and Felix. We were given the task of creating threaded PVC for the Hydroponics team. We have been cutting the PVC to length and then threading the unthreaded end at 14TPI. This was our first successful demonstration of the WISRD shop manufacturing abilities. The threads fit well into the receiving end for the hydroponics team. Overall, PVC is more forgiving then plastic, so it will be interesting to move to other materials like Aluminum. Our plan now is to create an instructional video on how to thread things using the lathe.
Week – February 13, 2017
This week I worked on the lathe, reinstalling the gear system. After this was completed we tested it out on a new project. We are assigned to create a set of threads matching 14 TPI (Threads per inch). We got some help from Marvin who is the chief machinist at UCLA. He showed us how to create threads and gave us some tips. On Wednesday we went to the Plasma Lab at UCLA and Marvin gave us some center drills, acid brushes, and some stock material. We figured out the correct gear assembly to achieve 14 TPI.
Week – February 6, 2017
This week I worked a lot with the lathe team on painting the metal rail piece with Rust-Oleum to make sure the welds don’t rust over time. The painting took some time because we had to wait a few minutes in between each coat before we could add another. Next, we plan to reinstall the drive gears so that the lathe can be fully operational.
Week – January 30, 2017
This week I focussed a lot on helping out around the lathe, helping to put the gear assembly back together. I then brainstormed with Felix, ideas about how we want to wire the “machining” area. At the moment we just have electrical wires stretching to the closest power bar. We are thinking of having a central outlet box as well as a key system to ensure nobody uses the machinery without permission.
Week – January 23, 2017
We changed the layout of the shop this week, moving things so that they all had their own area. I worked on wiring the computer for the CNC. I also helped Felix mount several monitors for the 3D printers. This gives much more work space on the counter tops for what ever is being worked on.
Week – January 16, 2017
This week we had a limited number of classes because of a day off, but I spent my time working on lathe. We sent in the broken gear rail to be welded back together, when it retuned that there was some material that needed to be filed down. We are still in progress filing this material down. Mean while, I reinstalled the bearing assembly for the feed and checked to see how it was working. I noticed some improvement from how it was working before. Not associated with WISRD, I got a chance to work with the 3D printers. I have never used them hand on but I already new a little bit about them. I made a print for another class and I learned a lot doing it.
Week – January 9, 2017
Throughout this week, I worked on the lathe with Blake and Felix. More specifically we were working on retrieving a key to repair a part that previously broke. We tried many methods of removing the key including, prying, dry ice, vice grips, and a propane torch. In the end we used a combination of heat from the propane torch and the vice grips to get it out. After being able to retrieve the part, we decided to take apart the bearing assembly for the feed. After removing two bolts, we cleaned inside of it and oiled it up. This made it significantly easier to turn.
Week – January 2, 2017
This week, I have worked with Felix and Blake on figuring out which supplies we will need to operate the Lathe. We looked at different tool holders that allow us to quickly swap out different cutting tools. We are still looking for something that will fit the mounting place on the lathe but should find something that works soon. In the shop, I have helped brainstorm with Blake and Felix about how we can make the shop structure better.