-
Transferable Deep Metric Learning for Clustering
Authors:
Simo Alami. C,
Rim Kaddah,
Jesse Read
Abstract:
Clustering in high dimension spaces is a difficult task; the usual distance metrics may no longer be appropriate under the curse of dimensionality. Indeed, the choice of the metric is crucial, and it is highly dependent on the dataset characteristics. However a single metric could be used to correctly perform clustering on multiple datasets of different domains. We propose to do so, providing a fr…
▽ More
Clustering in high dimension spaces is a difficult task; the usual distance metrics may no longer be appropriate under the curse of dimensionality. Indeed, the choice of the metric is crucial, and it is highly dependent on the dataset characteristics. However a single metric could be used to correctly perform clustering on multiple datasets of different domains. We propose to do so, providing a framework for learning a transferable metric. We show that we can learn a metric on a labelled dataset, then apply it to cluster a different dataset, using an embedding space that characterises a desired clustering in the generic sense. We learn and test such metrics on several datasets of variable complexity (synthetic, MNIST, SVHN, omniglot) and achieve results competitive with the state-of-the-art while using only a small number of labelled training datasets and shallow networks.
△ Less
Submitted 13 February, 2023;
originally announced February 2023.
-
Conv-NILM-Net, a causal and multi-appliance model for energy source separation
Authors:
Simo Alami C.,
Jérémie Decock,
Rim Kaddah,
Jesse Read
Abstract:
Non-Intrusive Load Monitoring (NILM) seeks to save energy by estimating individual appliance power usage from a single aggregate measurement. Deep neural networks have become increasingly popular in attempting to solve NILM problems. However most used models are used for Load Identification rather than online Source Separation. Among source separation models, most use a single-task learning approa…
▽ More
Non-Intrusive Load Monitoring (NILM) seeks to save energy by estimating individual appliance power usage from a single aggregate measurement. Deep neural networks have become increasingly popular in attempting to solve NILM problems. However most used models are used for Load Identification rather than online Source Separation. Among source separation models, most use a single-task learning approach in which a neural network is trained exclusively for each appliance. This strategy is computationally expensive and ignores the fact that multiple appliances can be active simultaneously and dependencies between them. The rest of models are not causal, which is important for real-time application. Inspired by Convtas-Net, a model for speech separation, we propose Conv-NILM-net, a fully convolutional framework for end-to-end NILM. Conv-NILM-net is a causal model for multi appliance source separation. Our model is tested on two real datasets REDD and UK-DALE and clearly outperforms the state of the art while kee** a significantly smaller size than the competing models.
△ Less
Submitted 13 February, 2023; v1 submitted 3 August, 2022;
originally announced August 2022.
-
CAMEO: Curiosity Augmented Metropolis for Exploratory Optimal Policies
Authors:
Simo Alami. C,
Fernando Llorente,
Rim Kaddah,
Luca Martino,
Jesse Read
Abstract:
Reinforcement Learning has drawn huge interest as a tool for solving optimal control problems. Solving a given problem (task or environment) involves converging towards an optimal policy. However, there might exist multiple optimal policies that can dramatically differ in their behaviour; for example, some may be faster than the others but at the expense of greater risk. We consider and study a di…
▽ More
Reinforcement Learning has drawn huge interest as a tool for solving optimal control problems. Solving a given problem (task or environment) involves converging towards an optimal policy. However, there might exist multiple optimal policies that can dramatically differ in their behaviour; for example, some may be faster than the others but at the expense of greater risk. We consider and study a distribution of optimal policies. We design a curiosity-augmented Metropolis algorithm (CAMEO), such that we can sample optimal policies, and such that these policies effectively adopt diverse behaviours, since this implies greater coverage of the different possible optimal policies. In experimental simulations we show that CAMEO indeed obtains policies that all solve classic control problems, and even in the challenging case of environments that provide sparse rewards. We further show that the different policies we sample present different risk profiles, corresponding to interesting practical applications in interpretability, and represents a first step towards learning the distribution of optimal policies itself.
△ Less
Submitted 15 February, 2023; v1 submitted 19 May, 2022;
originally announced May 2022.
-
A Study of Solar Irradiance Prediction Error Impact on a Home Energy Management System
Authors:
Mallek Mziou-Sallami,
Rim Kaddah,
Amira Ben Hamida
Abstract:
Nowadays, Energy Management Systems (EMS) are accessible for homes and buildings to optimize energy consumption especially when solar panels and batteries are installed. The intelligence of existing systems is often based on environmental or exogenous information like the weather, energy prices, and endogenous information like user consumption behavior and activity. The solutions aim to adapt a co…
▽ More
Nowadays, Energy Management Systems (EMS) are accessible for homes and buildings to optimize energy consumption especially when solar panels and batteries are installed. The intelligence of existing systems is often based on environmental or exogenous information like the weather, energy prices, and endogenous information like user consumption behavior and activity. The solutions aim to adapt a consumption profile to the produced energy in order to reduce costs. In the case of a perfect prediction of all variables, system performance can be controlled. In this article, we study the impact of generation prediction error on the daily energy cost. For this, we consider the energy management system as a black-box and we simulate multiple scenarios with different prediction errors using the quasi-random Monte Carlo method. We observe the global sensitivity of the system by measuring the Sobol indices in order to identify errors that impact more the daily energy cost. The analyses are based on French consumption data and on irradiance data for Carpentras, France. Results show that findings are aligned with battery charge and discharge strategies.
△ Less
Submitted 18 June, 2019;
originally announced June 2019.
-
Semi-Distributed Demand Response Solutions for Smart Homes
Authors:
Rim Kaddah,
Daniel Kofman,
Fabien Mathieu,
Michal Pioro
Abstract:
The Internet of Things (IoT) paradigm brings an opportunity for advanced Demand Response (DR) solutions. It enables visibility and control on the various appliances that may consume, store or generate energy within a home. It has been shown that a centralized control on the appliances of a set of households leads to efficient DR mechanisms; unfortunately, such solutions raise privacy and scalabili…
▽ More
The Internet of Things (IoT) paradigm brings an opportunity for advanced Demand Response (DR) solutions. It enables visibility and control on the various appliances that may consume, store or generate energy within a home. It has been shown that a centralized control on the appliances of a set of households leads to efficient DR mechanisms; unfortunately, such solutions raise privacy and scalability issues. In this chapter we propose an approach that deals with these issues. Specifically, we introduce a scalable two-levels control system where a centralized controller allocates power to each house on one side and, each household implements a DR local solution on the other side. A limited feedback to the centralized controller allows to enhance the performance with little impact on privacy. The solution is proposed for the general framework of capacity markets.
△ Less
Submitted 30 November, 2017;
originally announced November 2017.
-
Smart Fridge / Dumb Grid? Demand Dispatch for the Power Grid of 2020
Authors:
Joel Mathias,
Rim Kaddah,
Ana Bušić,
Sean Meyn
Abstract:
In discussions at the 2015 HICSS meeting, it was argued that loads can provide most of the ancillary services required today and in the future. Through load-level and grid-level control design, high-quality ancillary service for the grid is obtained without impacting quality of service delivered to the consumer. This approach to grid regulation is called demand dispatch: loads are providing servic…
▽ More
In discussions at the 2015 HICSS meeting, it was argued that loads can provide most of the ancillary services required today and in the future. Through load-level and grid-level control design, high-quality ancillary service for the grid is obtained without impacting quality of service delivered to the consumer. This approach to grid regulation is called demand dispatch: loads are providing service continuously and automatically, without consumer interference.
In this paper we ask, what intelligence is required at the grid-level? In particular, does the grid-operator require more than one-way communication to the loads? Our main conclusion: risk is not great in lower frequency ranges, e.g., PJM's RegA or BPA's balancing reserves. In particular, ancillary services from refrigerators and pool-pumps can be obtained successfully with only one-way communication. This requires intelligence at the loads, and much less intelligence at the grid level.
△ Less
Submitted 4 September, 2015;
originally announced September 2015.
