Princeton University Computer Science Department

Computer Science 598F Systems and Analytics for Big Data

Spring 2016

---

Directory

General Information | Syllabus | Projects

---

Tentative Schedule

Dates	Presenters	Topics	Readings	Presentations	To do
2/5	K. Li	General information
2/12	Guest lecture: Prof. Nick Turk-Browne, Princeton Neuroscience Institute	Real-time fMRI data analysis	Functional Interactions as Big Data in the Human Brain (Science, 2013) Closed-loop training of attention with real-time brain imaging (Nature-Neuroscience, 2015) Full correlation matrix analysis (FCMA): An unbiased method for task-related functional connectivity, (J. Neuroscience Methods, 2015) Full Correlation Matrix Analysis of fMRI Data on Intel Xeon Phi Coprocessors (SC, 2015)	pdf	Submit notes
2/19	K. Li Linpeng Tang	Some classic papers  Infrastructure Warmup exercise	MapReduce (OSDI, 2004),  GFS (SOSP, 2003), BigTable, (OSDI, 2006), Chubby (OSDI, 2006). Spark (NSDI, 2012) Spaark on a local cluster	pdf	Submit notes
2/26	Guest lecture: Prof. Sebastian Seung	Challenges in brain mapping	3D CNN for boundary detection (NIPS, 2015), A review paper (2012), Image segmentation challenge (ISBI, 2012)	link	Submit notes
3/4	Prof. Eric Xing (CMU) Dr. Zhifeng Chen (Google Brain team)	Petuum TensorFlow	Strategies and Principles of Distributed Machine Learning on Big Data (2015) Petuum: A New Platform for Distributed Machine Learning on Big Data (KDD 2015) TensorFlow: Large-scale machine learning on heterogeneous systems.   (Google white paper, 2015) Large Scaled Distributed Deep Networks.  (NIPS 2012).		Submit project proposal Submit notes
3/11	Yang, Bai	Dedup memory DNA data compression	HICAMP: Architectural Support for Efficient Concurrency-safe Shared Structured Data Access (ASPLOS, 2012) Log-Structured Memory for DRAM-based Storage.  (FAST 2014) HICAMP bitmap: space-efficient updatable bitmap index for in-memory databases (2014) A Survey of Techniques for Sequence Similarities Matching in Compression (2014) Data Compression for sequencing data (AMB 2013)		Submit notes
3/25	Chang, Naghib	Optimizations/High level features Arch support for Internet services	Making Sense of Performance in Data Analytics Frameworks (NSDI, 2015) Efficient Coflow Scheduling with Varys (SIGCOMM, 2014) Mastering the game of Go with deep neural networks and tree search.  (Nature 2016)		Submit notes Submit presentations
4/1	Kathpalia Guest lecture: Dr. Bill Tang (PPPL)	ImageNet competition Big Data in Fusion Research	ImageNet Classification with Deep Convolutional Neural Networks (NIPS, 2012). Deep Residual Learning for Image Recognition (2015) Big Data Machine Learning for Disruption Predictions White paper		Submit notes
4/8	Tang, Lin	Facebook social network	TAO: Facebook’s Distributed Data Store for the Social Graph (NSDI, 2013) Social Hash: An Assignment Framework for Optimizing Distributed Systems Operations on Social Networks (NSDI, 2016) Building Watson: An Overview of the DeepQA Project.  (AI Magazine, 2011).		Submit notes, Submit progress report
4/15	Qiu, Suo	Sparse FFT, Compressing DNA sequence data	Human genomes as email attachments (Bioinformationcs, 2009) The human genome contracts again (Bioniformatics, 2013), Simple and Practical Algorithm for Sparse Fourier Transform (SODA, 2012)		Submit notes
4/22	Zeng, Cheng	RGB-D Image data Clustering	Deep Sliding Shapes for Amodal 3D Object Detection in RGB-D Images (CVPR, 2016) Revisiting k-means: New Algorithms via Bayesian Nonparametrics Parallel Correlation Clustering on Big Graphs		Submit notes
4/29	Ko, Ravi	MOOC data CNN structure	Modeling and Prediction Learning Behavior in MOOCs (WSDM, 2016) Exploiting Cyclic Symmetry in Convolutional Neural Networks (ICML, 2016)		submit notes
5/13		Project final presentation			Submit final presentations
5/16		Final report submission			Submit final reports

Some Topics on Big Data Systems

• Storage Stack for Internet-Scale Systems
  Finding a needle in Haystack: Facebook's photo storage (OSDI, 2010)
  f4: Facebook's Warm BLOB Storage System. (OSDI, 2014)
  Pelican: A Building Block for Exascale Cold Data Storage.  (OSDI, 2014)
  
• Large-scale Database and Data warehouse
  F1: A Distributed SQL Database That Scales. (VLDB, 2013).
  Shark: SQL and Rich Analytics at Scale.  (SIGMOD 2013).
  Mesa: GeoReplicated, Near RealTime, Scalable Data Warehousing. (VLDB, 2014).
  A survey of large-scale analytical query processing in MapReduce.  (VLDB, 2014).
  Spanner: Google Globally Distributed Database (OSDI, 2012).
  
• Large-scale graph computing
  Pregel: a system for large-scale graph processing.  (SIGMOD, 2010).
  http://giraph.apache.org/Apache Giraph. 
  Distributed GraphLab: A framework for machine learning in the Cloud. (PVLDB, 2012).
  PowerGraph: Distributed graph-parallel computation on natural graphs.  (OSDI, 2012).
  GraphX: Graph Processing in a Distributed Dataflow Framework.  (OSDI, 2014)
  
  Facebook scaling Giraph.
  TAO: Facebook’s Distributed Data Store for the Social Graph (NSDI, 2013)
  Social Hash: An Assignment Framework for Optimizing Distributed Systems Operations on Social Networks (NSDI’2016)
  
• Architectural support for Internet services
  A Reconfigurable Fabric for Accelerating Large-Scale Datacenter Services (ISCA 2014).
  Achieving 10Gbps line-rate key-value stores with FPGAs.  (HotCloud, 2013).
  An FPGA-based In-line Accelerator for Memcached.  (Computer Architecture Letters, 2014).
• Memory Mechanisms for Large Datasets
  HICAMP: Architectural Support for Efficient Concurrency-safe Shared Structured Data Access (ASPLOS, 2012)
  Log-Structured Memory for DRAM-based Storage.  (FAST 2014)
• Achitectural Support for Deep Learning and Neurons
  DianNao: A Small-Footprint High-Throughput Accelerator for Ubiquitious Machine Learning.  (ASPLOS 2014)
  Efficient Digital Neurons for Large Scale Cortical Architectures (ISCA, 2014).
• FPGA Implementation of Convolutional Networks
  CNP: AN FPGA-BASED PROCESSOR FOR CONVOLUTIONAL NETWORKS. ( International Conference on FPGA, 2009)
  An FPGA-BAsed Stream Processor for Embedded Real-Time Vision with Convolutional networks.  (ECV 2009).
  Large-Scale FPGA-based Convolutional Networks.  (Machine Learning on Very Large Dataset, 2011).
  Optimizing FPGA-based Accelerator Design for Deep Convolutional Neural Networks (FPGA, 2015).
  

Some Topics on Analytics of Big Data

• Convolution Neural Networks
  Learning representations by back-propagating errors (Nature, 1985).
  Handwritten Digit Recognition with a Back-Propagation Network (NIPS, 1989).
  Visualizing and Understanding Convolutional Networks (ECCV, 2014)
  Deep learning (Nature, 2015)
  ZNN - Fast and Scalable Algorithm for 3D ConvNets on Multi-Core and Many-Core Shared Memory Machines (IPDPS 2016)
  
• ImageNet Competition Winners
  What does classifying more than 10,000 image categories tell us? (EECV 2009)
  2012 winner: ImageNet Classification with Deep Convolutional Neural Networks (NIPS, 2012).
  2014 winner: Going Deeper with Convolutions (CVPR, 2015)
  2015 winner: Deep Residual Learning for Image Recognition (2015)
  Current record: Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning (2016)
• Interesting AI Systems
  Mastering the game of Go with deep neural networks and tree search.  (Nature 2016)
  Building Watson: An Overview of the DeepQA Project.  (AI Magazine, 2011).
  
  
• Deep Learning on Large Datasets  (2 presentations)
  Building High-Level Features using Large Scale Unsupervised Learning.  (ICML 2012).
  Fast, Accurate Detection of 100,000 Object Classes on a Single Machine.  (CVPR 2013).
  Large-scale Video Classification with Convolutional Neural Networks.  (CVPR 2014).
  
  
• Clustering
  Parallel Correlation Clustering on Big Graphs (NIPS, 2015)
  Fast Distributed k-Center Clustering with Outliers on Massive Data (NIPS, 2015)
  Dependent nonparametric trees for dynamic hierarchical clustering (NIPS 2014)
  Dynamic Clustering via Asymptotics of the Dependent Dirichlet Process Mixture (NIPS 2013)