./20170404-0144-cet-state-of-the-art-23-lifelogging-presentation-personal-big-data-1.pdf

- The title for this presentation is "Lifelogging A New Big Data Challenge".

- Sample workflow of this project.
- I think mine is similar, but I would like to put sensors between the human and log.

- Extreme Lifelogging.
- This is like creating log (blog, vlog, ...) but done automatically.
- The lifelogging term means to capture real - life activities.
- The principle here is to use fitting wearable sensors.
- The result would be a personal awareness search engine that knows anything about you.
- However, this search engine is not intended to replace your organic memory.

- There is a slide that shows about the result of one year lifelogging.

- There is an assumption made here that the data captured from the lifelogging activities are all useful.
- I think this where machine learning could take in.


- This project is really about big data.
- There should be management mechanism somewhere.
- I think that management mechanism could be machine learning.
- At this point manual human editing is not feasible anymore.
- What is semantics?
- Semantics is a branch of languages that gives meaning.
- There are 2 main areas of semantics.
- Logical semantics, the implicit semantics that inferred from human senses.
- Lexical semantics, the explicit semantics.

- This is what I meant by context agent.
- Context agent answers all possible wh - based questions from values received from sensors.

- Example of application interface.
- I think the panel in the right shows you some usage of computer vision.
- However, I am not sure if computer vision can be that accurate.
- Perhaps, it is manual input by human.

- There is an example application developed within this "lifelogging" project titled My Visual Diary.
- This application shows images capture from the wide angle lens of the video sensor.

- 1 purpose of this "lifelogging" process is to have quantified self - analysis of one's personal life.

- Aside from personal use of traditional logging activities.
- There are other areas as well that could be benefited by having proper "lifelogging" tools. These are for examples.
- Depression management, Wittgenstein - esque.
- Healthcare.
- Lifestyle.

- The other purpose is to enhancing productivity.
- There should be a point in your life about "I know I have read that notion somewhere, somewhen.".
- This actually happens a lot to me during this project.

- "Lifelogging" could works both ways.
- It could works by logging 1 user.
- It could works by logging people.
- For example logging used by coffee robot to automatically determine which coffee each possible users would like.

- Example of personal memory search engine.
- There should be context and the agent to determine the relevancy of quantitative data.

- The concern is as usual, privacy.
- In this example third party face will be blurred.