Boston
The Boston chapter of Papers We Love
What was the last paper within the realm of computing you read and loved? What did it inspire you to build or tinker with? Come share the ideas in an awesome academic/research paper with fellow engineers, programmers, and paper-readers. Lead a session and show off code that you wrote that implements these ideas or just give us the lowdown about the paper. Otherwise, just come, listen, and discuss in a low ego, friendly environment.
Papers We Love has a Code of Conduct. Please contact one of the Meetup's organizers if anyone is not following it. Be good to each other and to the PWL community!
Chapter details
Sign-up: Please RSVP for meetings via Meetup.com
Twitter: @paperswelovebos
Organizers: Alex Benik, Richard Littauer
Chapter Meetups
MIT – Papers We Love Edition
For our February session of Papers We Love Boston, we’re partnering with MIT’s Classic Papers in Computer Science IAP seminar! Four attendees will present papers that were read and discussed as part of the seminar. We hope you can join us!
Speakers:
Anna Zeng, MIT CSAIL Graduate Student
"Enhancing Server Availability and Security Through, Failure-Oblivious Computing”,
Martin Rinard, Cristian Cadar, Daniel Dumitran, Daniel M. Roy, Tudor Leu, and William S. Beebee, Jr
Cynthia Zeng, MIT Operations Research Graduate Student
"'A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence”,
McCarthy, Minsky, Rochester, Shannon
Aubrey Simonson, MIT Media Lab Graduate Student
"Moving Objects In Space: Exploiting Proprioception In Virtual-Environment Interaction”, Mark Mine
Robert Solomon, MIT Alumnus
“As We May Think”, The Atlantic Monthly, 1945, Vannevar Bush
Scott Sanderson — Frank McSherry's COST and Timely Dataflow
"Scalability! But at What COST?" and "Naiad: A Timely Dataflow System"
In this talk, we'll look at two papers, "Scalability! But at What COST?" and "Naiad: A Timely Dataflow System". Many of the observations and arguments from "Scalability! But at What COST?" were motivated by the authors' experience of working on Naiad. Naiad's timely dataflow model has since been ported to Rust, where it sees active development.
Scalability! But at What COST?
Distributed data processing systems often aim to provide "scalability", offering users the ability to accelerate computations by taking advantage of many machines. But building a scalable system introduces significant overheads, and many systems force programmers into restricted programming models that make it hard to use efficient algorithms. In "Scalability! But at What COST?", McSherry et al. argue that researchers should measure the COST (Configuration that Outperforms a Single Thread) of their scalable systems.
<…The Best Refactoring You've Never Heard Of (James Koppel)
We're excited to have James Koppel reprise his talk from this summer's Compose Conference NYC!
Doors will open at 6pm, talk will start at 6:30.
Every problem can be solved by many designs. Often the relation between two designs is that one is a transformation of another. When we understand how to mechanically transform designs, we can more easily explore the design space, and effortlessly identify the tradeoffs.
So, what do these design changes have in common?
- Letting a search procedure take an arbitrary filter function, instead of a fixed set of options
- Changing a program using blocking I/O to non-blocking I/O with an event loop
- Letting a user stop in the middle of an action — and resume it after a server reboot
Answer: They’re all instances of a transformation called “defunctionalization” or its inverse, “refunctionalization.” Come learn how it works, and learn to recognize when a proposed design change is just an instance of …
Jacob Lisi on Gorilla, a fast, scalable in-memory timeseries database
Paper: Pelkonen, Franklin, Teller, Cavallaro, Huang, Meza, Veeraraghavan. Gorilla: A Fast, Scalable, In-Memory Time Series Database. VLDB 2015.
…Steve Willis — Advertising Entanglement Capabilities in Quantum Networking
Background for the Internet Draft "Advertising Entanglement Capabilities in Quantum Networks”
Quantum communications - the exchange of quantum information between remote systems - is evolving from simple point-to-point communication to a general purpose multi-hop network.
Developing this extended capability introduces a new spin to a familiar set of networking problems :
What’s unique about quantum information transfer?
How does quantum mechanics hinder and help the transfer of quantum information?
How do you build a general purpose network that supports quantum information transfer?
I will try to explain, to an audience familiar with networking, how entanglement, teleportation, and swapping might be combined with classical IP networking and routing to build a quantum network.
From the draft abstract : "This document describes the use of link-state routing protocols on classical links in Quantum Networks. It contains proposals for additions…
Yifan Xing – Consensus algorithm, Paxos and Raft
Many messaging systems that are widely used in the industry, e.g., Kafka, use centralized distributed systems services to achieve reliability and consensus between servers. Companies in the industry use the services; however, only a few of them understand the details of the protocols. This talk brings the principles used in academia to the industry by introducing the common distributed systems protocols implemented underneath the popular services. In addition, this talk will compare the differences between how the protocols are used in both academia and the industry. It provides details of how the protocols, specifically Paxos and Raft, work, including how they elect leaders among servers, how they achieve consensus between machines, and how they reliability process and execute client commands. Therefore, it shows how the systems and services, which use the protocols, are enabled to have fault-tolerance, and to achieve confidentiality, integrity, authenticity, availability, etc. Fro…
Douglas Creager on Pastry peer-to-peer overlay network
The early 2000's saw several research groups look into distributed hash tables as a decentralized way to structure a distributed system. In this talk, I'll briefly describe some of the history behind DHTs, and look at one of them in particular: Pastry, which takes into account how close or far apart nodes are from each other when constructing the peer-to-peer overlay network. We'll also look at Scribe, which implements anycast/multicast messaging on top of the Pastry overlay network.
Papers:
A. Rowstron and P. Druschel, "Pastry: Scalable, distributed object location and routing for large-scale peer-to-peer systems". IFIP/ACM International Conference on Distributed Systems Platforms (Middleware), Heidelberg, Germany, pages[masked], November, 2001.
A. Rowstron, A-M. Kermarrec, M. Castro and P. Druschel, "SCRIBE: The design of a large-scale event notification infrastructure", NGC2001, UCL, London, November 2001.
Niko Matsakis on Lambda Prolog
Niko Matsakis, one of the core Rust contributors, is going to present some proof techniques from Prolog that are informing some advanced work in Rust's type checker and compiler.
Traditional Prolog lets you define logical relationships and then have the solver figure out the inferences. But Prolog's Horn Clauses are fairly simplistic in terms of what they can express. This paper talks about a few relatively simple techniques that lets us extend to a much richer logic, supporting things like proving that one statement implies another, or proving "for all" predicates. So for example where Prolog could perhaps setup rules to decide when `hungry(joe)` is true or `hungry(nancy)`, this technique describes techniques to prove `hungry(P)` is true for *any* person P.
"A proof procedure for the logic of hereditary Harrop formulas", Gopalan Nadathur, Journal of Automated Reasoning, Feb 1993, 11(1), pp[masked].
What happened to distributed programming languages? (Heather Miller)
We're very excited to have Heather Miller reprise her PWLConf talk!
Abstract
Nowadays, most programs we write are in some sense distributed—making HTTP requests or serving responses over HTTP, fetching or computing data on some remote resource, building some microservice that is meant to interact with others, etc. With all of this distribution going on, one might ask, what happened to distributed programming languages? Why are we still using languages like Java or C++ for these sorts of tasks? In this talk, Heather will take us on a whirlwind tour through history up to the present of distributed programming languages as well as programming constructs meant for distribution like futures and RPC. Together, we'll try to work out what happened to all of the distributed programming languages!
References Programming Models for Distributed Comp…
DNA computing (Chelsea Voss)
Join us as Chelsea Voss presents a handful of papers on DNA computing:
• Leonard Adleman (the "A" in RSA). Computing with DNA.
• Faulhammer et al. Molecular computation: RNA solutions to chess problems.
• Qian + Winfree's Strand displacement circuits
• Rothemund + Winfree + Papadakis, Algorithmic Self-Assembly of DNA Sierpinski Triangles – "DNA origami" implementing cellular automata
…Implementing lazy functional languages on stock hardware (Joe Jevnik)
Join us for some programming language pragmatics! Joe Jevnik will present Simon Peyton Jone's "Implementing lazy functional languages on stock hardware: the Spineless Tagless G-machine".
Paper abstract:
The Spineless Tagless G-machine is an abstract machine designed to support nonstrict higher-order functional languages. This presentation of the machine falls into three parts. Firstly, we give a general discussion of the design issues involved in implementing non-strict functional languages.
Next, we present the STG language, an austere but recognisably-functional language, which as well as a denotational meaning has a well-defi ned operational semantics. The STG language is the "abstract machine code" for the Spineless Tagless G-machine.
Lastly, we discuss the mapping of the STG language onto stock hardware. The success o…
Godel's Incompleteness Theorems
Godel's Incompleteness Theorems are among the most famous results in modern mathematics. They are often informally stated as follows:
1. Any formal system that can express elementary arithmetic contains statements that are "true but not provable".
2. Any consistent formal system that can express elementary arithmetic cannot prove its own consistency.
These informal statements, however, often raise more questions than they answer. What exactly is a "formal system"? What does it mean for a statement to be "true but not provable"? What does it mean to "express arithmetic", and why does arithmetic have anything to do with proof or consistency?
This talk will introduce the audience to the main ideas behind the incompleteness theorems. We'll discuss the historical context in which the incompleteness theorems were discovered, develop the concepts necessary to properly understand what the theorems actually say, and present a sketch of the theorems' proof.
…You are not expected to understand this - Arun Thomas on Unix Process Switching
"You are not expected to understand this." That is the now famous comment in UNIX v6's process switching (swtch) code. Process switching -- and multitasking more generally -- is a neat magic trick the OS pulls off. It's a delicate dance between the operating system and the hardware. We will dig into how this dance works on UNIX v6 and the PDP11, including the famous comment. We will also look into how the process switching magic happens on a more recent OS and CPU architecture (FreeBSD on x86). It should be a fun trip down the rabbit hole of operating system kernels internals and instruction set specifications
…An oversimplified history of CSP -Talk by Douglas Creager
An oversimplified history of CSP
Communicating Sequential Processes (CSP) was first described by Tony Hoare in 1978. Since then, there have been two main "branches" of further development. In the first branch, CSP inspired how concurrency is handled in an ongoing series of programming languages — most recently (and famously) in Go. In the second branch, CSP was refined into a rigorous "process calculus", and has since been used to formally prove various safety and liveness properties of concurrent and distributed systems.
This talk will be a whirlwind history of CSP, a high-level overview of the process calculus, and some live demos of using a refinement checker to prove things about interesting systems. (No proving things by hand!)
CAR Hoare. "Communicating Sequential Processes", CACM 21(8), pp. [masked], 1978.
CAR Hoare. Communi…
Jonathan Schuster on Programmatic and Direct Manipulation, Together at Last
Have you ever wished GUI tools for building visuals like PowerPoint and Illustrator had a more programming-like interface for making similar images rather than copying, pasting, and tweaking all the time? Have you ever wished programmatic systems like LaTeX and Processing.js allowed you to just click and drag on items in the output to change the result, rather than going through an edit-compile-run cycle every time? In some of the latest research, you can have both!
In this talk, Jon Schuster will discuss "Programmatic and Direct Manipulation, Together at Last" [1] from this year's PLDI (Conference on Programming Language Design and Implementation). The paper proposes a method to automatically edit an SVG-generating program to match the changes made by dragging components of the resulting image around the canvas, while still allowing the developer to edit the code directly. A prototype is available online [2], so expect interactive demos!
A shorter position paper [3] t…
Tim Riser on Mastering the Game of Go with Deep Neural Networks & Tree Search
Tim Riser on Mastering the Game of Go with Deep Neural Networks & Tree Search
A grand challenge of AI has fallen - a decade earlier than the ever-abused experts predicted. But should we care?
What made AlphaGo, the AI built by DeepMind, so unique?
To dive into AlphaGo's system of learning, evaluation, and search algorithms that combined to defeat the reigning Go world champion, Tim will be presenting on "Mastering the Game of Go with Deep Neural Networks & Tree Search", a paper by Google DeepMind.
http://airesearch.com/wp-content/uploads/2016/01/deepmind-mastering-go.pdf
Bio:
Tim Riser is currently researching at the Berkman Center for Internet & Society at Harvard University. He graduated from Brigham Young University wit…
Steve Willis on Xerox Network Systems (XNS)
What we gonna do right now is go back.
How far back?
Way back.
Intro- You probably know that many influential technologies were developed by Xerox: Ethernet, networking protocols, bitmapped workstations, laser printing technology and object oriented programing, etc.
So brush up on your OSI model and be ready to dive into networking archeology. Come here the story of XNS, which became the canonical protocol in early days of local area networking.
XNS greatly influenced protocols such as Novell NetWare, and Banyan VINES, AppleTalk and lives on within the corners of the Internet. https://goo.gl/Ev3atn
BIO -
Steven Willis is an engineer, executive and co-founder of networking companies such as Wellfleet and Argon. From 1982 to 1986, he worked as a developer at Interlan, one of the original Ethernet controller companies, developing …
Douglas Creager on The Early History Of Smalltalk by Alan C. Kay
History! Those who don't know it are doomed to etc.
Smalltalk is one of the earliest object-oriented languages. It's extremely influential — many of the OO concepts that we use in today's languages were first developed in Smalltalk. But what influenced Smalltalk? Alan Kay will walk us through the language's inspirations in this talk, including such crazy new technologies as UIs with overlapping windows and a 1970's-era ur-tablet!
http://worrydream.com/EarlyHistoryOfSmalltalk/
Bio:
Douglas Creager is a software engineer at Google with an eclectic history. He got a CS Master's from MIT working in a biology group, and a PhD from Oxford while teaching formal methods to professional software engineers. He's worked as a consultant helping finance and telecoms firms wrangle data between different systems, and was VP of Engineering for a network security startup based out of DC. He liv…
Peter Bailis on Streaming Queries Over Streaming Data
PWL Boston / Cambridge
We are thrilled to have Peter Bailis as the speaker for our April meeting. Peter is brilliant researcher and outstanding speaker. He will be presenting on Streaming Queries over Streaming Data" from VLDB 2002 by Sirish Chandrasekaran and Michael J. Franklin.
If you are looking for a preview of the type of talks Peter gives, check out this talk given at the PWL SF chapter in late 2014 on Bayou
Bio
After spending this year visiting MIT CSAIL, Peter Bailis will join Stanford Computer Science as an assistant professor. Peter's research in the Future Data Systems group (http://futuredata.stanford.edu) focuses on the design and…
Richard Littauer on The Moral Character of Cryptographic Work
PWL Boston / Cambridge is back!
Ripped from the headlines, but not Law and Order, Richard Littauer will be presenting on "The Moral Character of Cryptographic Work" by Phillip Rogaway of UC Davis.
Bio
"Richard Littauer is a computer scientist and developer who lives in Cambridge. He works for IPFS (ipfs.io), building a distributed, permanent internet, and reads papers and books in his constantly dwindling spare time. He's most well known for writing the dictionary for the Na'vi language, and for getting paid to review websites while drunk at theuserisdrunk.com."
Logistics
PWL Boston / Cambridge is hosted at the Battery Ventures office in the Seaport:
One Marina Park Drive, FL 11
Boston, MA 02210
Upon arriving at One Marina Park Drive, you will need your IDs to check in…