Kien Trung's PhD defence

Last week I had the pleasure of being an examiner for Nguyen Kien Trung's PhD thesis at the Université Nice Sophia Antipolis. Kien's thesis focused on a an efficient electricity disaggregation algorithm which he deployed using a low-cost system-on-chip. Kien successfully defended his thesis against the panel's questions, which sometimes came in a mixture of English and French, which I found particularly very impressive!

I also managed a brief visit to Qualisteo's office, who demonstrated some the disaggregation algorithms they're applying across a range of built environments, including the Eiffel Tower, a motorway tunnel and a sports stadium. I had a great time, but hopefully my next trip to Nice will last a little longer than 24 hours!

NILM @ GlobalSIP 2015

The 3rd IEEE Global Conference on Signal and Information Processing will be held next month (14-16 December 2015) in Orlando, Florida. I was excited to see quite how many papers in the program are related to energy disaggregation, which are split across two sessions organised as part of the Smart Buildings Symposium and the Inference and Prediction session:

Toward a Semi-Supervised Non-Intrusive Load Monitoring System for Event-based Energy Disaggregation. Karim Said Barsim and Bin Yang (University of Stuttgart, Germany)

A new approach for supervised power disaggregation by using a deep recurrent LSTM network. Lukas Mauch and Bin Yang (University of Stuttgart, Germany)

Blind Non-intrusive Appliance Load Monitoring using Graph-based Signal Processing. Bochao Zhao, Vladimir Stankovic and Lina Stankovic (University of Strathclyde, United Kingdom)

A New Unsupervised Event Detector For Non-Intrusive Load Monitoring. Benjamin Wild, Karim Said Barsim and Bin Yang (University of Stuttgart, Germany)

Dataport and NILMTK: A Building data set Designed for Non-intrusive Load Monitoring. Oliver Parson (University of Southampton, United Kingdom); Grant Fisher and April Hersey (Pecan Street Inc, USA); Nipun Batra (IIIT Delhi, India); Jack Kelly (Imperial College London, United Kingdom); Amarjeet Singh (IIIT-Delhi, India); William J Knottenbelt (Imperial College London, United Kingdom); Alex Rogers (University of Southampton, United Kingdom)

Non-Intrusive Load Monitoring: A Power Consumption Based Relaxation. Kyle Anderson, Jose Moura and Mario Berges (Carnegie Mellon University, USA)

A feasibility study of automated plug-load identification from high-frequency measurements. Jingkun Gao (Carnegie Mellon University, USA); Emre C Kara (Lawrence Berkeley National Laboratory, USA); Suman Giri and Mario Berges (Carnegie Mellon University, USA)

Single-Channel Compressive Sampling of Electrical Data for Non-Intrusive Load Monitoring. Michelle Clark and Lutz Lampe (University of British Columbia, Canada)

Non-Intrusive Load Monitoring of HVAC Components using Signal Unmixing. Alireza Rahimpour (The University of Tennessee at Knoxville, USA); Hairong Qi (the University of Tennessee, USA)

I'm really looking forward to the conference, and will do my best to update this post with links to papers as they become available.

Hangout On Air: NILMTK Algorithm Interface Discussion

This Friday, we're planning to have a discussion around how best to integrate existing disaggregation algorithms into NILMTK via a Google Hangout On Air. The Hangout will take place at 16.30 UTC on Friday 20th November 2015, which I believe translates to 08.30 PST and 22.00 IST. Jack Kelly, Nipun Batra and myself will be available give the NILMTK perspective, while Stephen Makonin and Mingjun Zhong will be joining us to discuss how they'd like to integrate their algorithms into the toolkit.

A provisional agenda for the call is:

• Overview of NILMTK algorithm interface - Oliver Parson (15 mins)
• Discussion of sparseNILM algorithm - Stephen Makonin (15 mins)
• Discussion of Latent Bayesian Melding applied to NILM - Mingjun Zhong (15 mins)
• General Q&A (15 mins)

I've created a Google+ event which provides more details about the call. If you'd like to join the discussion, please follow the link to the event and launch the Hangout from the event page. If you'd prefer just listen in rather than joining the discussion, you can watch and listen live on YouTube. The video from the Hangout should be available on YouTube after the call, and I will try to update this post with the corresponding link.

Update 23.11.2015: a video from the call is available on YouTube.

BuildSys 2015: My Disaggregation is Better Than Yours

Akshay presenting his paper on LocED

Last week I had the pleaure of chairing a session named 'My disaggregation is better than yours', at the 2015 BuildSys conference in Seoul. The session featured presentations of the following three papers:

• Multi-User Energy Consumption Monitoring and Anomaly Detection with Partial Context Information. Pandarasamy Arjunan (IIIT-Delhi), Harshad D Khadilkar (IBM Research), Tanuja Ganu (IBM Research), Zainul M Charbiwala (IBM Research), Amarjeet Singh (IIIT-Delhi), Pushpendra Singh (IIIT-Delhi)
• LocED: Location-aware Energy Disaggregation Framework. S.N. Akshay Uttama Nambi (TUDelft), Antonio Reyes Lua (TUDelft), R. Venkatesha Prasad (TUDelft)
• Neural NILM: Deep Neural Networks Applied to Energy Disaggregation. Jack Kelly (Imperial College London), William Knottenbelt (Imperial College London) [joint best presentation]

Even outside this session, there were a few other papers relevant to NILM, including the following papers:

• If You Measure It, Can You Improve It? Exploring The Value of Energy Disaggregation. Nipun Batra (IIIT-Delhi), Amarjeet Singh (IIIT-Delhi), Kamin Whitehouse (University of Virginia)
• Health Monitoring of Elderly Residents via Disaggregated Smart Meter Data and Log Gaussian Cox Processes. Jose Alcala (University of Alcala), Oliver Parson (University of Southampton), Alex Rogers (University of South Hampton)
• Contextual Air Conditioning Disaggregation with Probabilistic Soft Logic. Sabina Tomkins (University of California Santa Cruz), Lise Getoor (University of California Santa Cruz)

To the best of my knowlege, this is the first conference with published proceedings with such a large amount of work related energy disaggregation. I'm glad that NILM is starting to find a common home at such a great venue, and I'm already looking forward to BuildSys 2016 at Stanford!

Energy disaggregation for health monitoring

José and I have been working on a project to apply NILM methods to the health monitoring domain, specifically to help monitor the activity of elderly people living independently in their own homes. The approach we pursued aims to disaggregate the kettle reliably from smart meter data, without requiring any training in each home. We chose to use the kettle as the appliance of interest as it is an appliance common to almost all UK homes, is used regularly as part of most elderly people’s daily routine, and also has a signature which varies little between houses.

The core novelty of the our work is that deviations in kettle usage from the normal routine can be recognised from the disaggregated smart meter data, allowing interventions in households to occur as soon as possible. Crucially, such routines are learned individually for each household, rather than using a static routine for all households. As such, kettle usage is used as a proxy for health, since no additional sensors were installed to directly measure health parameters (e.g. heart rate).

An example of a routine for a single household is shown below. The top graph shows the probability that the kettle would be used at least once within a half hour interval, while the bottom graph shows the cumulative probability that the kettle would have been used at least once by that time of the day. It can be seen that for this household, a highly repeatable routine is followed, where the kettle is used by midday on about 80% of days.

Full details can be found in our paper to be presented at BuildSys later this year:

José Alcalá, Oliver Parson, Alex Rogers. Detecting Anomalies in Activities of Daily Living of Elderly Residents via Energy Disaggregation and Cox Processes. In: 2nd ACM International Conference on Embedded Systems For Energy-Efficient Built Environments (BuildSys), Seoul, South Korea. 2015.

Dataport data released in NILMTK format

The Dataport database is the world's largest source of disaggregated customer energy data. The database contains electricity data collected from 722 houses in the US; 631 in Texas, 49 in Colorado and 42 in California. The houses monitored include 501 single-family homes, 183 apartments, 35 town homes and 3 mobile homes. Access to the portal is free for members of universities, while commercial access is limited to members of Pecan Street's Industry Advisory Council.

To date, the Dataport data has been available via direct access to the database. While this provides an efficient means of querying a small amount of data, large amounts of data can take a long amount of time to download since the data is transferred in an uncompressed format. Furthermore, as with most other data sets, the data set is described in a custom format, requiring researchers to parse the data and metadata before making use of the data set.

For these reasons, we've been working with Pecan Street Inc to release a subset of the Dataport database in NILMTK HDF5 format. The HDF5 file is 1.09 GB in size, and contains one month of data from 669 of the Dataport houses, which were selected as they contain at least 8 meters. In each house, the circuit name has been converted from the Dataport names to the NILM Metadata controlled vocabulary. The produced dataset can be easily analysed using tools described in the NILMTK documentation. The HDF5 file is available via the Dataport portal under the same access control as the Dataport database.

Below is a boxplot showing the proportion of energy consumed by each circuit category across all households in the HDF5 data.

The data set is described in more detail in the following paper to be presented at the GlobalSIP Smart Buildings workshop:

Oliver Parson, Grant Fisher, April Hersey, Nipun Batra, Jack Kelly, Amarjeet Singh, William Knottenbelt, Alex Rogers. Dataport and NILMTK: A Building Data Set Designed for Non-intrusive Load Monitoring. In: 1st International Symposium on Signal Processing Applications in Smart Buildings at 3rd IEEE Global Conference on Signal & Information Processing, Orlando, FL, USA, 14-16 December, 2015.

Reusable hold out test sets for NILM

Overfitting is a well-cited problem in the field of Non-intrusive Load Monitoring. Overfitting refers to the high accuracy of an algorithm on one small data set, while the same algorithm generalises poorly to other data sets. In NILM, this often corresponds to algorithms which work well on data covering a short period of time or a small number of houses, but performs poorly on data covering longer time periods or other houses.

One solution to this problem is through the use of a competition, in which the organiser releases part of the data set for training, while holding out the remaining data for testing. This approach works fine for one-off competitions where each participant can only submit one solution, but it weakens when participants are allowed to submit multiple entries. The reason for this is that participants can use information learned from their accuracy score on the test set to inform their algorithm choice.

Belkin organised exactly this type of competition for NILM via the Kaggle platform a few years ago. The competition allowed two entries per day per participant, each of were evaluated using half of a private hold out data set and displayed on a public leaderboard. However, the competition ended after 4 months, at which point the final standings were determined by each solution's performance on the other half of the private hold out data set which had never previously been released. Through such a format, the final standings can only be calculated once and the competition cannot be re-run, as the final standings convey information about the private hold out data set which could inform the design of future algorithms.

A recent Google Research blog post describes this exact problem in a much more general sense than NILM. Most interestingly, their recent paper even proposes a solution to this problem through the use of a reusable holdout set. The approach is that the reusable hold out set is only accessed through a differentially private algorithm, which effectively samples the holdout set in order to produce a different sample each time it is accessed.

At the 2015 European NILM workshop, an MSc group from Imperial supervised by Jack Kelly presented a platform which could potentially be used to host NILM competitions in the future. I’d be really interested to see whether such a platform could use such a reusable hold out set in order to allow the competition to run for much longer without compromising the results relative to the classical method of evaluation.