Computer Laboratory

MPhil, Part III, and Part II Project Suggestions (2017-2018)

Contact

 

 

 

 

 

 

 

  

 

Please contact Eiko Yoneki (email: eiko.yoneki@cl.cam.ac.uk) if you are interested in any project below.

1.  Dynamic Task Scheduling on Heterogeneous CPU/GPU Environment using ML for Parallel Processing

Contact: Eiko Yoneki

In this project, various aspects of parallel processing will be explored using a new generation of CPU/GPU integrated board. We use NVIDIA’s Jetson X2. This new GPUs makes it possible to cluster the GPU nodes for different scale of parallel processing. Various applications running over X2 will be investigated including graph processing and neural network training using the Tensorflow framework. Graph processing can take advantage of processor heterogeneity to adapt to structural data patterns. The overall aim of graph processing can be seen as scheduling irregular tasks to optimise data-parallel heterogeneous graph processing, by analysing the graph at runtime and dispatching graph elements to appropriate computation. Various task scheduling strategies over NN applications could be explored such as mixture of model parallelism and data parallelism.

In comparison to the above approach, several types of GPU machines in cluster computing environment will be explored. Furthermore, our recent work, BOAT [1], could be used for optimisation of complex parameter space.

[1] V. Dalibard, M. Schaarschmidt, and E. Yoneki: BOAT: Building Auto-Tuners with Structured Bayesian Optimization, WWW, 2017.  https://github.com/VDalibard/BOAT.

2.  Distributed Stochastic Gradient Descent on Naiad

Contact: Eiko Yoneki

Naiad is a distributed system framework that allows for automatic parallelisation of programs in a distributed environment [1]. In this project, you will use the RUST implementation of NAIAD and build a distributed implementation of Stochastic Gradient Descent (SGD) [2], which is a common algorithm used in Machine Learning. The project will provide efficient parallel processing for SGD in optimised fashion, where different iterations of the algorithm for different data points can be run in parallel. An efficient distributed implementation using NAIAD can take advantage of the incremental and iterative computation. You would also write an application to evaluate the project.

[1] D. Murray, F. McSherry, R. Isaacs, M. Isard, P. Barham, and M. Abadi. Naiad: a timely dataflow system. SOSP, 2013.

[2] Kevin P Murphy. Machine learning: a probabilistic perspective. MIT press, 2012.

3.  Adaptive Stream Processing with Deep Reinforcement Learning

Contact: Eiko Yoneki (with Michael Schaarschmidt)

Stream processing engines such as Apache Storm or Heron enable real-time distributed event processing. Recently, the concept of self-regulating stream processing has been introduced to help stream processing engines adapt to new loads and deal with failures [1]. In prior open source work, this has been achieved via a set of manually defined rules connecting certain events in the cluster (e.g. higher load) with certain actions (add nodes to cluster). The goal of this project is to investigate whether this system can be improved by learning an end-to-end control mechanism via reinforcement learning and foregoing manual adjustment rules.

A library of robust reinforcement learning algorithms is available as a starting point [2]. The key research question in this project is to understand how real time performance metrics in a distributed system can be leveraged in a reinforcement learning setting. This project requires strong programming ability in Python/Java, and an interest to learn about emerging developments in machine learning/reinforcement learning.

[1] https://www.microsoft.com/en-us/research/wp-content/uploads/2017/06/p1218-floratou.pdf

[2] https://github.com/reinforceio/tensorforce

4. Optimisation of JVM Garbage Collection using ML

Contact: Eiko Yoneki

This project extends the Cassandra case study [1,2] in incremental ways:

1)     Cassandra database was only executed in one setting. We could create an auto-tuner that is able to deal with a wider variety of settings. This includes:

 

-        Changing the heap size of the JVM

-        Changing the underlying hardware of Cassandra

-        Changing the underlying hardware of YCSB (Yahoo! Cloud Serving Benchmark)

 

2)     Applying it to other JVM applications (such as other databases, anything where we have a rigorous benchmark for throughput and latency)

 

3)     Looking at some of the other JVM flags (we only tuned three) as well as the other JVM garbage collector

For evaluation following two points are evaluated.

1)     No GC option as baseline

2)     Various GC implementations (i.e. some are good on certain workloads)

[1] http://www.cl.cam.ac.uk/~ey204/pubs/2017_WWW.pdf

[2] http://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-900.html

5. Optimising Newral Network Model over Power Consumption

Contact: Eiko Yoneki

In this project, you would tune neural network model itself. Pick one typical Deep Neural Network application (e.g. image classification), and play with various hyper-parameters, where hyper-parameters might take substantially different values for the best accuracy or performance. Especially you can set an objective function on reducing the power consumption.

You would explore various parameters based experiments and build up a probabilistic model expressing the parameter relationship applying the BOAT technique (see [1] and [2]).

[1] http://www.cl.cam.ac.uk/~ey204/pubs/2017_WWW.pdf

[2] http://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-900.html

6. Building Graph Query Function using Functional Programming

Contact: Eiko Yoneki

Demand to store and search of online data with graph structure is emerging. Such data range from online social networks to web links which requires efficient query processing in a scalable manner. In this project, you will build a graph query function (layer) to achieve efficient graph data processing. The graph query function builds on a lazy graph loaded from multiple sources and performs queries at the same speed or faster than a relational database. The function should be written in Clojure or another functional language to support practical lazy loading from data source to an in-memory graph database. Efficient representations for graph queries should be also investigated. The evaluation includes comparisons with existing Graph Database and the query capability comparing with SQL. You could start the project by our existing work in this area.

References:

[1] Microsoft Research, “Trinity project: Distributed graph database,” http://research.microsoft.com/en-us/projects/trinity/

[2] Neo Technology, “Java graph database,” http://neo4j.org/

7. Unikernel over Raspberry Pi: Building Data Processing Platform

Contact: Eiko Yoneki

This project builds a unikernel-based distributed data processing platform over Raspberry Pi clusters, where mapreduce type of operation or simple neural network training and inference can be run. A unikernel is a form of monolithic operating system, but instead of being general purpose a unikernel is built with a single task in mind. By removing the unneeded functionality of an operating system, unikernels bring about a variety of advantages such as extremely small compared to conventional operating systems. Such small footprint makes movement of code to the distributed nodes much faster and easier to deal with limited availability of memory in Raspberry pi.

We pursue a project, RaDiCS [2], where unikernel currently runs over Raspberry pi using QEMU emulator and it can be a starting point of the project. One of the challenges in this project is optimising the performance of unikernel in Raspberry pi. Potential applications running over the built distributed platform will be a type of neural network applications, where each feature can be trained separately without constant synchronisation over the layer of NN model, and parallelism using 100s of Paspberry pis can be deployed.

References:

[1] Jitsu: Just-In-Time Summoning of Unikernels, NSDI, 2015.

[2] RaDiCS https://gitlab.com/RaspiUnikernel/RaDiCS.

8. Approximate Algorithms Determining Local Clustering Coefficients Anonymously

Originator/Supervisor: Eiko Yoneki (with Amitabha Roy in Intel Labs)

Keywords: Sampling, Approximation, Privacy, Cluster Coefficient

Anonymous social networks are a new phenomenon in an increasingly privacy conscious world. A natural question to ask in this setting is whether we can continue to apply known principles of network science in such settings where neighbourhood information is concealed both between nodes and external observers. This project is to work on approximate algorithms that determines clustering coefficient in such settings. Clustering coefficients provide a way to relatively order nodes by importance in a network, determine densely connected neighbourhoods and distinguishing social graphs from web graphs. Algorithms to measure clustering coefficients have hitherto required global knowledge of graph links. This requires individual nodes in the graph to give up the identity of their neighbours. This project investigates an algorithm for estimating the clustering coefficient of a vertex by exchanging only anonymised set summaries of neighbours, from which it is difficult to reverse engineer individual links. The bulk of the project will consist of working on and improving sampling techniques to arrive at accurate local clustering coefficients without exchanging explicit neighbour lists in social networks .

[1] P. Flajolet, Eric Fusy, O. Gandouet, and et al. Hyperloglog: The analysis of a near-optimal cardinality estimation algorithm. In Proceedings of the International Conference on Analysis of Algorithms, 2007.

9. RasPi-Net: Building Stream Data Processing Platform over RasPiNET

Originator/Supervisor: Eiko Yoneki

Keywords: Raspberry Pi, Delay Tolerant Networks, Satellite Communication, Stream Processing

We have built a decentralised Raspberry Pi network (RasPiNET [1]), which can be deployed in wild and remote regions as a standalone network. The gateway Raspberry Pi nodes are integrated with satellite communication devices, where the light version of Delay Tolerant Network (DTN) bundle protocol is embedded. RasPiNET could consist of 10-100 nodes. As an example, a remote sensing application could be written either in RasPi or Smart phones that can connect to RasPi. Collected data could be processed within RasPiNET to reduce data size that streams over the satellite communication to the base location. The crowd sourcing application can run on top of RasPiNET, too. The goal of this project is building a stream processing platform in both directions: from data collection from RasPiNET nodes to the data processing nodes possibly via a satellite gateway and from bulk of data delivery to the satellite gateway node to disseminate necessary information to RasPiNET nodes. A good filtering function and RasPiNET in-network data aggregation could be developed.

References:

[1] E. Yoneki: RasPiNET: Decentralised Communication and Sensing Platform with Satellite Connectivity.  ACM CHANTS, 2014.

[2] Delay Tolerant Network Bundle Protocol: http://tools.ietf.org/html/rfc6255

[3] RockBlock Technology:http://rockblock.rock7mobile.com/

10. Clustering Entities across Multiple Documents in Massive Scale

Originator/Supervisor: Eiko Yoneki

Keywords: Clustering, Graph Partitioning, Random Walk, Distributed Algorithms

Many large-scale distributed problems including the optimal storage of large sets of graph structured data over several hosts - a key problem in today’s Cloud infrastructure. However, in very large-scale distributed scenarios, state-of-the-art algorithms are not directly applicable, because frequent global operations over the entire graph are difficult. In [1], balanced graph partitioning is achieved by a fully distributed an algorithm, called Ja-be-Ja, that uses local search and simulated annealing techniques for graph partitioning annealing techniques for graph partitioning. The algorithm is massively parallel: each node is processed independently, and only the direct neighbours of the nodes and a small subset of random nodes in the graph need to be known locally. Strict synchronisation is not required. These features allow Ja-be-Ja to be easily adapted to any distributed graph processing system. This project starts understanding Ja-be-Ja as a starting point, and investigates further performance improvement. A case study: a graph-based approach to coreference resolution, where a graph representation of the documents and their context is used and applying a community detection algorithm based in [1] can speed up the task of coreference resolution by a very large degree.

[1] Fatemeh Rahimian, Amir H. Payberah, Sarunas Girdzijauskas, Mark Jelasity and Seif Haridi: JabeJa: A Distributed Algorithm for Balanced Graph Partitioning, IEEE International Conference on Self-Adaptive and Self-Organizing Systems (SASO), 2013.

[2] Fatemeh Rahimian, Amir H. Payberah, Sarunas Girdzijauskas, and Seif Haridi: Distributed Vertex-Cut Partitioning, DAIS, 2014.

Contact Email

Please email to eiko.yoneki@cl.cam.ac.uk for any question.