Derek Rayside presents current research on the field of microfluidics. Microfluidics are currently developed mainly by trial and error. How can this be improved?
Microfluidics is an exciting new area concerned with designing devices that perform some medical diagnoses and chemical synthesis tasks orders of magnitude faster and less expensively than traditional techniques. However, microfluidic device design is currently a black art, akin to how digital circuits were designed before 1980.
We have developed a hardware description language that is appropriate for the description and synthesis of both single-phase and multi-phase microfluidic devices. These are new results that have not yet been published. This is collaborative work with other research groups in Mechanical Engineering, Chemical Engineering, and Electrical Engineering.
GM for the W2015 term, two main amendments to be discussed: Requiring elections to be held within two weeks of the beginning of term and adopting a club-wide code of conduct. <br> Code Party 1 follows, we're doing timed code golf problems, T-shirts might find themselves on people who do well on code golf.
GM for the W2015 term, two main amendments to be discussed: Requiring elections to be held within two weeks of the beginning of term and adopting a club-wide code of conduct.
Code Party 1 follows, we're doing timed code golf problems, T-shirts might find themselves on people who do well on code golf.
Day 2 of Runtime Type Inference in Dynamic Languages with Kannan Vijayan
Day 2 of Runtime Type Inference in Dynamic Languages with Kannan Vijayan
Javascript is fast. In some cases, very close to compiled-language fast. How is this even possible? How do we know what types our variables have? How can we optimize it well? Kannan Vijayan will be talking about the historical advances in JIT-compilation of dynamically typed programs over two days. Of course, both of those talks will have free food.
How do we make dynamic languages fast? Today, modern Javascript engines have demonstrated that programs written in dynamically typed scripting lan- guages can be executed close to the speed of programs written in languages with static types. So how did we get here? How do we extract precious type information from programs at runtime? If any variable can hold a value of any type, then how can we optimize well?
This talk covers a bit of the history of the techniques used in this space, and tries to summarize, in broad strokes, how those techniques come together to enable efficient jit-compilation of dynamically typed programs. To do the topic justice, Kannan Vijayan will be talking the Monday and Tuesday March 9th and 10th.
Does that mean two consecutive days of free food? Yes it does.
Murphy Berzish explains how to programmatically determine if a program is satisfiable, and how to find a concrete counterexample if it is unsatisfiable. At the core are SAT/SMT solvers. SAT theory deals with Boolean Satisfiability solvers, while SMT theory--Satisfiability Modulo a Theory--allows SMT to be extended to common data structures. Free food!
Does your program have an overflow error? Will it work with all inputs? How do you know for sure? Test cases are the bread and butter of resilient design, but bugs still sneak into software. What if we could prove our programs are error-free?
Boolean Satisfiability (SAT) solvers determine the ‘satisfiability’ of boolean set of equations for a set of inputs. An SMT solver (Satisfiability Modulo a Theory) applies SMT to bit-vectors, strings, arrays, and more. Together, we can reduce a program and prove it is satisfiable, or provide a concrete counter-example. The implications of this are computer-aided reasoning tools for error-checking in addition to much more robust programs.
In this talk Murphy Berzish will give an overview of SAT/SMT theory and some real-world solution methods. He will also demonstrate applications of SAT/SMT solvers in theorem proving, model checking, and program verification.
What else? Oh yes, refreshments and drinks will be served. Come out!
The first code party of Winter 2015, and we have something a litle different this time. We're running a Code Retreat (coderetreat.org) with Boltmade. The result of this is that you will be able to do a coding challenge, wherein you implement Rule 110 (like the Game of Life). Of course, if you want to work on whatever you can do that as well. Delicious free food, but RSVP! [bit.ly/code-party-0](<https://bit.ly/code-party-0>)
The first code party of Winter 2015, and we have something a litle different this time. We're running a Code Retreat (coderetreat.org) with Boltmade. The result of this is that you will be able to do a coding challenge, wherein you implement Rule 110 (like the Game of Life). Of course, if you want to work on whatever you can do that as well. Delicious free food, but RSVP! bit.ly/code-party-0
Part of the Cheriton School of CS' Distinguished Lecture Series, MIT's Leslie Kaelbling will discuss robotic AI applied to the messy real world. We make a number of approximations during planning but regain robustness and effectiveness through a continuous state estimation and replanning process. This allows us to solve problems that would otherwise be intractable to solve optimally.
The fields of AI and robotics have made great improvements in many individual subfields, including in motion planning, symbolic planning, probabilistic reasoning, perception, and learning. Our goal is to develop an integrated approach to solving very large problems that are hopelessly intractable to solve optimally. We make a number of approximations during planning, including serializing subtasks, factoring distributions, and determinizing stochastic dynamics, but regain robustness and effectiveness through a continuous state estimation and replanning process. This approach is demonstrated in three robotic domains, each of which integrates perception, estimation, planning, and manipulation.
Theo Belaire, a fourth-year CS student, will be talking about Racket's match' function. Bug resistant, legible, and super powerful! Especially useful for CS 241 in writing compilers, but all-round a joy to write.
Come learn how to use the power of the Racket match construct to make your code easier to read, less bug-prone and overall more awesome!
Theo Belaire, a fourth-year CS student, will show you the basics of how this amazing function works, and help you get your feet wet with some code examples and advanced use cases.
If you're interested in knowing about the more powerful features of Racket, then this is the talk for you! The material covered is especially useful for students in CS 241 who are writing their compiler in Racket, or are just curious about what that might look like.
Alex Tsay from AeroFS will talk about the high availability distributed file systems they develop.The CAP Theorem outlined the fundamental limitations of a distributed system. When designing a distributed system, one has to constantly be aware of the trade-off between consistency and availability. Most distributed systems are designed with consistency in mind. However, AeroFS has decided to build a high-availability file system instead. In this tech talk, I'll be presenting an overview of AeroFS file system, advantages and challenges of a high-availability file system, and examine the inner workings of AeroFS's core syncing algorithm.
Alex Tsay from AeroFS will talk about the high availability distributed file systems they develop.
The CAP Theorem outlined the fundamental limitations of a distributed system. When designing a distributed system, one has to constantly be aware of the trade-off between consistency and availability. Most distributed systems are designed with consistency in mind. However, AeroFS has decided to build a high-availability file system instead. In this tech talk, I'll be presenting an overview of AeroFS file system, advantages and challenges of a high-availability file system, and examine the inner workings of AeroFS's core syncing algorithm.
Elections for Winter 2015 are being held! Submit a nomination and join your fellow members in choosing this term's CSC executive. (Please note the time change to 7PM.)
The Computer Science Club will be holding its termly elections this upcoming Thursday, Jan. 15 at 6PM in the Comfy Lounge (MC 3001). During the election, the president, vice-president, treasurer and secretary will be elected, the sysadmin will be ratified, and the librarian and office manager will be appointed.
Nominations are now closed. The candidates are:
President:- Luke Franceschini (l3france
)
Gianni Gambetti (glgambet
)
Ford Peprah (hkpeprah
)
Khashayar Pourdeilami (kpourdei
)
Vice-President:- Luke Franceschini (l3france
)
Gianni Gambetti (glgambet
)
Patrick Melanson (pj2melan
)
Ford Peprah (hkpeprah
)
Khashayar Pourdeilami (kpourdei
)
Treasurer:- Weitian Ding (wt2ding
)
Aishwarya Gupta (a72gupta
)
Edward Lee (e45lee
)
Secretary:- Ilia "itchy" Chtcherbakov (ischtche
)
Luke Franceschini (l3france
)
Patrick Melanson (pj2melan
)
Ford Peprah (hkpeprah
)
Khashayar Pourdeilami (kpourdei
)
Voting will be heads-down, hands-up, restricted to MathSoc social members. If you'd like to review the elections procedure, you can visit our Constitution page.
Our speaker, Avery Pennarun, will share some not-very-secret secrets from the team creating GFiber's open source router firmware, including some discussion of wifi, marketing truthiness, the laws of physics, something about coaxial cables, embedded ARM processors, queuing theory, signal processing, hardware design, and kernel driver optimization. If you're lucky, he may also rant about poor garbage collector implementations. Also, there will be at least one slide containing one of those swooshy circle-and-arrow lifecycle diagrams, we promise.Please RSVP here: http://bit.ly/GoogleFiberTalk.
Google Fiber's Internet service offers 1000 Mbps internet to a few cities: that's 100x faster than a typical home connection. The problem with going so fast is it moves the bottleneck around: for the first time, your Internet link may be faster than your computer, your wifi, or even your home LAN.
Our speaker, Avery Pennarun, will share some not-very-secret secrets from the team creating GFiber's open source router firmware, including some discussion of wifi, marketing truthiness, the laws of physics, something about coaxial cables, embedded ARM processors, queuing theory, signal processing, hardware design, and kernel driver optimization. If you're lucky, he may also rant about poor garbage collector implementations. Also, there will be at least one slide containing one of those swooshy circle-and-arrow lifecycle diagrams, we promise.
About Avery Pennarun: Avery graduated from the University of Waterloo in Computer Engineering, started some startups and some open source projects, and now works at Google Fiber on a small team building super fast wifi routers, TV settop boxes, and the firmware that runs on them. He lives in New York.
Please RSVP here: http://bit.ly/GoogleFiberTalk.