Axinia Radeva

Machine Learning for the New York City Power Grid

Power companies can benefit from the use of knowledge discovery methods and statistical machine learning for preventive maintenance. We introduce a general process for transforming historical electrical grid data into models that aim to predict the risk of failures for components and systems. These models can be used directly by power companies to assist with prioritization of maintenance and repair work. Specialized versions of this process are used to produce (1) feeder failure rankings, (2) cable, joint, terminator, and transformer rankings, (3) feeder Mean Time Between Failure (MTBF) estimates, and (4) manhole events vulnerability rankings. The process in its most general form can handle diverse, noisy, sources that are historical (static), semi-real-time, or real-time, incorporates state-of-the-art machine learning algorithms for prioritization (supervised ranking or MTBF), and includes an evaluation of results via cross-validation and blind test. Above and beyond the ranked lists and MTBF estimates are business management interfaces that allow the prediction capability to be integrated directly into corporate planning and decision support; such interfaces rely on several important properties of our general modeling approach: that machine learning features are meaningful to domain experts, that the processing of data is transparent, and that prediction results are accurate enough to support sound decision making. We discuss the challenges in working with historical electrical grid data that were not designed for predictive purposes. The “rawness” of these data contrasts with the accuracy of the statistical models that can be obtained from the process; these models are sufficiently accurate to assist in maintaining New York Citys electrical grid.

A process for predicting manhole events in Manhattan

We present a knowledge discovery and data mining process developed as part of the Columbia/Con Edison project on manhole event prediction. This process can assist with real-world prioritization problems that involve raw data in the form of noisy documents requiring significant amounts of pre-processing. The documents are linked to a set of instances to be ranked according to prediction criteria. In the case of manhole event prediction, which is a new application for machine learning, the goal is to rank the electrical grid structures in Manhattan (manholes and service boxes) according to their vulnerability to serious manhole events such as fires, explosions and smoking manholes. Our ranking results are currently being used to help prioritize repair work on the Manhattan electrical grid.

21st-Century Data Miners Meet 19th-Century Electrical Cables

Electrical grid reliabil ity will be a key issue as peak demand for electric ity continues to increase. Grids will need to accommodate a growing population, more high-tech appliances, distributed power genera tion, and, soon, a large fleet of electric vehicles.

Reducing Noise in Labels and Features for a Real World Dataset: Application of NLP Corpus Annotation Methods

This paper illustrates how a combination of information extraction, machine learning, and NLP corpus annotation practice was applied to a problem of ranking vulnerability of structures (service boxes, manholes) in the Manhattan electrical grid. By adapting NLP corpus annotation methods to the task of knowledge transfer from domain experts, we compensated for the lack of operational definitions of components of the model, such as serious event . The machine learning depended on the ticket classes, but it was not the end goal. Rather, our rule-based document classification determines both the labels of examples and their feature representations. Changes in our classification of events led to improvements in our model, as reflected in the AUC scores for the full ranked list of over 51K structures. The improvements for the very top of the ranked list, which is of most importance for prioritizing work on the electrical grid, affected one in every four or five structures.

Analytics for Power Grid Distribution Reliability in New York City

We summarize the first major effort to use analytics for preemptive maintenance and repair of an electrical distribution network. This is a large-scale multiyear effort between scientists and students at Columbia University and the Massachusetts Institute of Technology and engineers from the Consolidated Edison Company of New York Con Edison, which operates the worlds oldest and largest underground electrical system. Con Edisons preemptive maintenance programs are less than a decade old and are made more effective with the use of analytics developing alongside them. Some of the data we used for our projects are historical records dating as far back as the 1880s, and some of the data are free-text documents typed by Con Edison dispatchers. The operational goals of this work are to assist with Con Edisons preemptive inspection and repair program and its vented-cover replacement program. This has a continuing impact on the public safety, operating costs, and reliability of electrical service in New York City.

Cross-lingual sentiment transfer with limited resources

We describe two transfer approaches for building sentiment analysis systems without having gold labeled data in the target language. Unlike previous work that is focused on using only English as the source language and a small number of target languages, we use multiple source languages to learn a more robust sentiment transfer model for 16 languages from different language families. Our approaches explore the potential of using an annotation projection approach and a direct transfer approach using cross-lingual word representations and neural networks. Whereas most previous work relies on machine translation, we show that we can build cross-lingual sentiment analysis systems without machine translation or even high quality parallel data. We have conducted experiments assessing the availability of different resources such as in-domain parallel data, out-of-domain parallel data, and in-domain comparable data. Our experiments show that we can build a robust transfer system whose performance can in some cases approach that of a supervised system.

Prediction of a hotspot pattern in keyword search results

Abstract This paper identifies and models a phenomenon observed across low-resource languages in keyword search results from speech retrieval systems where the speech recognition has high error rate, due to very limited training data. High confidence correct detections ( hccd s) of keywords are rare, yet often succeed one another closely in time. We refer to these close sequences of hccd s as keyword hotspots . The ability to predict keyword hotspots could support speech retrieval, and provide new insights into the behavior of speech recognition systems. We treat hotspot prediction as a binary classification task on all word-sized time intervals in an audio file of a telephone conversation, using prosodic features as predictors. Rare events that follow this pattern are often modeled as a self-exciting point process ( sepp ), meaning the occurrence of a rare event excites a following one. To label successive points in time as occurring within a hotspot or not, we fit a sepp function to the distribution of hccd s in the keyword search output. Two major learning challenges are that the size of the positive class is very small, and the training and test data have dissimilar distributions. To address these challenges, we develop a novel data selection framework that chooses training data with good generalization properties. Results exhibit superior generalization performance.

Multivariate Assessment of a Repair Program for a New York City Electrical Grid

We assess the impact of an inspection repair program administered to the secondary electrical grid in New York City. The question of interest is whether repairs reduce the incidence of future events that cause service disruptions ranging from minor to serious ones. A key challenge in defining treatment and control groups in the absence of a randomized experiment involved an inherent bias in selection of electrical structures to be inspected in a given year. To compensate for the bias, we construct separate models for each year of the propensity for a structure to have an inspection repair. The propensity models account for differences across years in the structures that get inspected. To model the treatment outcome, we use a statistical approach based on the additive effects of many weak learners. Our results indicate that inspection repairs are more beneficial earlier in the five-year inspection cycle, which accords with the inherent bias to inspect structures in earlier years that are known to have problems.