|
Princeton
University
|
Computer Science 598C
|
Spring 2013 |
Directory
General Information | Syllabus | Projects
Tentative Syllabus
(You can
find all presentations on this course’s blackboard)
Week 1
2/4: Organizational meet (digital universe)
2/8: MapReduce, datasets and project ideas, (Google paper, Dean’s keynote,
Chapter 2 of MMDS, critic blog which has been removed)
Warm-up exercise:
Follow these instructions to access the C8 cluster. Modify the given WordCount example (source code is at HDFS:/user/fuse/WordCount.java) such that your code will clean up the corpus documents (HDFS:/user/fuse/essays) and output the frequencies of words in a sorted order. Your final program will be submitted to department dropbox (we will setup shortly). Due: 2/22.
Week 2 (2/15): Deep learning and GFS
Linpeng Tang (reading review, secondary high-level structure, ICLM’12)
Sharvanath Pathak (reading GFS, SOSP’03, secondary Chubby, OSDI’06)
submit reading nodes
Week 3 (2/22):
Guest lecture by Dr. Phillip Shilane from EMC (reading acceleration for backup, FAST’12)
Brian Tubergen (reading protocol independent dedup, SIGCOMM’00, secondary network dedup acceleration)
submit reading nodes
submit warmup exercise
2/24
Submit
project proposals
Week 4 (3/1):
Guest lecture by Dr. Ruoming Pang from Google (reading Google’s globally distributed DB, OSDI’12)
Mehmet Basbug
(primary: LSH (VLDB’99), secondary Muti-probe (VLDB’07)).
Guest: Dr. William Hanson (CMO of UPenn Hospital) on medical data and projects
Submit reading notes
Week 5 (3/8):
Madhuvanthi Jayakumar (Primary: Google’s globally distributed storage OSDI’10)
Nayden Nedev (Primary: Google Data center network, CACM’12, SDN)
Christian Edbank (Primary: KNN-Decent, WWW’11, Secondary: 7.1-7.3 of MMDS,)
Submit reading notes
Week 6 (3/15)
Akshay Mittal (primary: Ramcloud, secondary: Spark)
Srinivas Narayana (primary: MMDS 4.1-4.4, secondary Opensketch)
Andrew Werner (Primary: Spanner OSDI’12, secondary Paxos)
Submit reading notes
Week 7: (3/29)
Guest lecture by Dr. Sanjeev Kumar from Facebook (reading Facebook’s photo storage, OSDI’10)
Xiao Li (Primary: Amazon’s key-value store SOSP’07, secondary SILT SOSP’11).
Submit reading notes
Project mid-term
progress reports and short presentations,
submit
Week 8 (4/5)
Eric First (Primary: MMDS 3.5, Secondary: graph similarity)
Wathsala Wathawana (Primary: layered naming, Secondary: Chord)
Muneeb Ali (Primary: Kahn’s paper, Secondary: Lampson’s paper)
Submit reading notes
Week 9 (4/12)
Mike Mckeown (Primary: energy-proportional computing; Secondary: energy-efficient mapreduce)
Trevor Bannard (Primary: ImageNet construction; Secondary: classification with 10,000 categories)
Robert Sami (Primary: 9.1-9.2 MMDS book; Secondary: 9.3 MMDS book)
Submit reading notes
Week 10 (4/19)
Sachin Ravi (Primary: external hashing, secondary: unpublished)
Marcela Melara (Primary: info leak in cloud; secondary: attack-resource)
Tri Nguyen (Primary: energy-proportional storage; secondary: manycore key-value store)
Submit reading notes
Week 11 (4/26)
Guest lecture by Prof. Moses Charikar (various topics)
Alp Kutlualp (primary: chapter 11.1 and 11.2, secondary: chapter 11.3http://i.stanford.edu/~ullman/pub/ch11.pdf)
Submit reading notes
Last class meet (5/17)
All: Project demos/presentations (each 15 minutes)
5/19
Project report due, Submit
Suggested book
Mining of Massive Data Sets. Anand Rajaraman, Jure Leskovec, and Jeffrey D. Ullman. Cambridge University Press. 2011.
You can download the latest book from an author’s webpage.
Systems Topics
MapReduce abstraction
Google paper, Dean’s keynote, Chapter 2 of MMDS, critic blog (removed but interesting for discussion)
Google systems
GFS (SOSP’03), BigTable, (OSDI’06), Google cluster(IEEE’03), Google Data center network (CACM’12),
Key-value stores
Amazon’s key-value store (SOSP’07), SILT (SOSP’11)
Distributed storage systems
Facebook’s photo storage (OSDI’10), Google’s globally distributed storage (OSDI’10), Google’s globally distributed DB (OSDI’12), Microsoft’s Azure storage (USENIX’12).
Deduplication storage systems
Venti (FAST’02), Data domain DDFS (FAST’08), others (TBD).
Deduplication WAN bandwidth optimization
WAN protocol independent dedup (SIGCOMM’00), LBFS (SOSP’01), WAN acceleration for backup (FAST’12), others (TBD).
Analytics topics
Domain specific feature extractions
Image SIFS (ICCV’99)), Audio MFCC ( ), others (TBD)
Unsupervised feature learning and deep learning
review, high-level structure (ICLM’12), others (TBD)
Ontology
Imagenet (CVPR’09, ECCV’10), others (TBD)
Similarity measures
Section 3.5 of MMDS, papers (TBD)
Shingles and minhashing
Section 3.2-3.3 of MMDS, Fingerprinting, Document resemblance
Locality sensitive hashing
Section 3.4 of MMDS, LSH (VLDB’99), Muti-probe (VLDB’07), Posteriori Multi-probe, …
Dimension reduction
Curse of Dimensionality (TBD), PCA (TBD), Sketches (TBD), others (TBD)
Streaming
Sampling (Section 4.1 and 4.2 of MMDS), Bloom filter (Section 4.3 of MMDS), Counting (section 4.4 of MMDS), others (TBD)
Processing large data in memory
Section 6.3 of MMDS
Clustering in high Dimensional Space
K-means (Section 7.3 of MMDS), KNN-Decent (WWW’11), others (TBD)
Web link analysis
Parts of Chapter 5 of MMDS
Graph search
Parts of chapter 10 of MMDS