Robust-first Computing Wiki - people:xinyu_chen

abstractv2

anonymous@undisclosed.example.com (Anonymous) — Sun, 12 Oct 2014 21:28:16 +0000

@@ -1 +1,10 @@
+ ====== The Simulation of Interactions Between Big Structures and Small Particles on The Movable Feast Machine ======
  
+ 
+ ===== Abstract =====
+ 
+ Too much medicine is harmful. Can we increase the efficiency of medicine? This paper will simulate the interaction between cells and medicine particles on the Movable Feast Machine. The simulation will include the process of growth and replication of cells, as well as the movements of medicine particles. This attempt is inspired by some basic biological phenomena. We will try to study the self-replicating and density-control behaviors of big structures like cells. These studies can help us to do jobs outside the event-window. Maybe we can increase the robustness by doing so. 
+ 
+ ===== Hoped-for Outcome =====
+ 
+ Basic elements are Cell and Medicine. A group of cell atoms will form a big structure to resemble a real Cell. Medicine atoms move randomly so they only have certain possibilities to find tumors and kill them. To increase the efficiency of medicine particles, we designed an element called Aura. Tumors will emit Aura atoms so they attract medicine particles. In real world, this Aura can be implemented by magnetic fields or electric fields. However too strong aura will also damage healthy cells.  We will adjust the threshold of Aura to find out its effect on the whole situation.

abstractv3

anonymous@undisclosed.example.com (Anonymous) — Mon, 20 Oct 2014 19:37:42 +0000

@@ -1 +1,10 @@
+ ====== How Small cures Big: The Simulation of interactions between Medicine and Tumor on The Movable Feast Machine ======
  
+ 
+ ===== Abstract =====
+ 
+ This paper reports the simulation of interactions between cells and medicine on the Movable Feast Machine. This attempt is inspired by some basic biological phenomena. The simulation applies a virtual electric field method to affect the movements of medicine particles. So that it can increase the probabilities for medicines to enter Tumor cells. We also studied the self-replicating and density-control behaviors of big structures like cells. These studies can help us to accomplish jobs outside the event-window thus can increase the robustness of the Movable Feast Machine. 
+ 
+ ===== Hoped-for Outcome =====
+ 
+ Basic elements include Cell, Medicine, M-Emitter and M-Consumer. A group of cell atoms forms a big structure to resemble a real Cell. Medicine atoms move randomly. They only have certain possibilities to find tumors and kill them. To increase this curing chances, we designed the Aura element. Tumors will emit Aura atoms which can be attract Medicine atoms. In real world, this Aura can be implemented by magnetic fields or electric fields. Auras have a timer representing their life time. The longer this timer is, the strong the virtual magnetic fields is. The effect of attracting medicine is stronger then. However, too strong auras also damage healthy cells.  We will adjust the threshold of the Aura strength to find out its effect on the whole system.

abstractv4

anonymous@undisclosed.example.com (Anonymous) — Tue, 04 Nov 2014 21:46:58 +0000

@@ -1,10 +1,10 @@
- ====== Maybe This Way: An Adaptive Stochastic One-Way Communication in Distributed Systems ======
+ ====== Maybe This Way: Adaptive Stochastic One-Way Communications in Distributed Systems ======
  
  
  ===== Abstract =====
  
- Inspired by some imaginary biological phenomena,this paper experiments the long distance communication in distributed systems by adaptive stochastic attempts. The experiments let tumors emit virtual smells to attract medicine particles. Those virtual smells act as one-way messengers to help roving medicines to locate tumors. On the Movable Feast Machine, communication within event-windows is accurate and bidirectional. Without unique global addresses, these experiments help to extend communication distances outside event-windows, thus to increase the robustness of the Movable Feast Machine. 
+ Inspired by biology, we present a model of long distance communications in a distributed system. The experiments let tumors emit virtual smells to attract medicine particles. Those virtual smells act as one-way messengers to help roving medicines to locate tumors. On the Movable Feast Machine, communication within event-windows is accurate and bidirectional. Without unique global addresses, these experiments help to extend communication distances outside event-windows, thus to increase the robustness of the Movable Feast Machine. 
  
- ===== Hoped-for Outcome =====
+ ===== Introduction =====
  
- Basic elements include Cell, Medicine, M-Emitter and M-Consumer. A group of cell atoms forms a big structure to resemble a real Cell. Medicine atoms move randomly. They only have certain possibilities to find tumors and kill them. To increase this curing chances, we designed the Aura element. Tumors will emit Aura atoms which can be attract Medicine atoms. In real world, this Aura can be implemented by magnetic fields or electric fields. Auras have a timer representing their life time. The longer this timer is, the strong the virtual magnetic fields is. The effect of attracting medicine is stronger then. However, too strong auras also damage healthy cells.  We will adjust the threshold of the Aura strength to find out its effect on the whole system.  
+ Basic elements include Tumor, Medicine, and Aura. When Medicines roaming the universe, they only have a certain possibility to find Tumors and kill them. To increase the chances for Medicines to locate Tumors, we let Tumors to emit Auras as if they are sending out messengers to Medicines. In the world of distributed computation, Tumors can be some hosts in need of certain resources, and Auras can be broadcast messages. We can use these messages to allocate resources without a global address manage mechanism. This is a step towards more robust computations.

drugmissile

anonymous@undisclosed.example.com (Anonymous) — Sun, 12 Oct 2014 21:25:01 +0000

@@ -1,9 +1,12 @@
- ===== Drug missiles: Dynamic Medicine Route Simulation: =====
+ ===== Drug attractors: Dynamic Medicine Route Simulation: =====
  
  When people are ill, they take tablets, capsules or get injections. What happens to medicines once they are in our bodies? How do they find the right place? Are they chasing virus or bacteria like the shark-fish game? 
  
- Although medicines take different routes and they fight different illness like pain, infection or cancer, they normally will move into the blood stream and then organs and tissues. I would like to simplify this process and focus on a particular stage. Assume the drug molecules are at the gate of the infected organ. The project will first simulate how these molecules find their targets now. The next step will imagine how they can be guided to their targets in the future by magnet or electric charge forces. 
+ Although medicines take different routes and they fight different illness like pain, infection or cancer, they normally will move into the blood stream and then organs and tissues. I would like to simplify this process and focus on a particular stage. Assume the drug particles have already entered the ill organ. The project will simulate the  growth and reproduction of cells. Then imagine medicine can be attracted by 'Aura', some magnet or electric charge forces, to increase their efficiency. 
+ 
+ [[people:xinyu_chen:abstractV2|Title and Abstract]]    
+ 
  ==== Basic Elements included here are: ====
  === Medicine_Particles: ===
    * small(its proportion to cells as 1 to 10^3);
    * very large amount(10^4)
@@ -19,17 +22,25 @@
    * if the group size>= x(like 5), group will stop(v=0) and swallowed by leukocyte 
  === Cells: ===
    * volume are large, occupy many spaces(30*30, about one tile??)
    * stay still, do not move
-   * grow, divide when volume is big enough(like vl>=1500)
-   * have membrane
+   * grow, divide at certain time  
        * Good_cells: grows after some time(like t1=10^3)
        * Tumor_cells:grows faster than good cells(like t2<=500);
            * by a probability(like pt=10%) turns good cells into tumor by a time(like t3>= 200)
            * killed by particles(like n>=10)
+           * emit 'Aura'
+ === Aura: ===
+   * small atoms, diffuse around
+   * have a short life-time(explode after a short time)
+   * attract medicine particle(like fish be eaten by shark)
+   * density will be high around tumors(because they die very soon)
+   * will do harm to healthy cells(if they live long enough, they will 'eat' healty cell atoms)
+   * The life-length and number of Aura reflect how strong it is.
+  
  === Blood_Streams: ===
    * Particles will move along blood streams. 
    * Empty spaces must be left between Cells. 
  ==== Big Challenge: ====
    * Create appropriate Cells is very difficult. Cells must have reasonable shapes, they should not look like square rooms.
    * Cells can grow and divide. They will bump away any particles within that event window.
    * When Cells grow and divide, they will not squeeze out empty spaces. There are always some spaces for Blood streams.

wgw

anonymous@undisclosed.example.com (Anonymous) — Sun, 26 Oct 2014 20:23:28 +0000

@@ -21,5 +21,8 @@
  {{ :people:xinyu_chen:maxresdefault.jpg?300|}}In 1951, Dr. W.Grey Walter displayed his `tortoises' at the Festival of Britain. His electric toy simulated two basic characteristics of animal behavior:goal-seeking and scanning. The designer-endowed goal was moderate illumination. One vacuum tube was used to simulate two interconnected neurons. The first sensor was a photocell and it was connected to the drive and steering motors. The second sensor was a contact switch that indicated that the turtle’s “shell” had bumped into something. The simple actions of stop, back up ,turn allowed the turtles to move around or avoid obstacles. They could wander a room and return to a hutch to recharge batteries. His two papers, "An Imitation of Life" in 1950, "A Machine that Learns" in 1951 and his book "The Living Brain" in 1963 were widely read and studied.
  
  The technique he used are reflected in today’s biologically-inspired robots. These are the B.E.A.M philosophy. The acronym BEAM is standing for: Biology, Electronics, Aesthetics, Mechanics. Biology means we look the nature for inspirations, to solve our problems; Electronics is the technology for us to get our creations work; Aesthetics means that something "looks cool" rather than something crappy though they may work; Mechanics is an intelligent design such as your robot can get around without optical or touch switches.  
  
-  
+ ==== Papers and Books ====
+   *{{:people:xinyu_chen:walter50imitation.pdf|An Imitation of Life(1950)}}
+   *{{:people:xinyu_chen:walter51learns.pdf|A Machine that Learns(1951)}}
+   *The Living Brains(1963)