Pankaj Malhotra

Diksha Garg, Pankaj Malhotra, Anil Bhatia et al.

We consider learning a trading agent acting on behalf of the treasury of a firm earning revenue in a foreign currency (FC) and incurring expenses in the home currency (HC). The goal of the agent is to maximize the expected HC at the end of the trading episode by deciding to hold or sell the FC at each time step in the trading episode. We pose this as an optimization problem, and consider a broad spectrum of approaches with the learning component ranging from supervised to imitation to reinforcement learning. We observe that most of the approaches considered struggle to improve upon simple heuristic baselines. We identify two key aspects of the problem that render standard solutions ineffective - i) while good forecasts of future FX rates can be highly effective in guiding good decisions, forecasting FX rates is difficult, and erroneous estimates tend to degrade the performance of trading agents instead of improving it, ii) the inherent non-stationary nature of FX rates renders a fixed decision-threshold highly ineffective. To address these problems, we propose a novel supervised learning approach that learns to forecast the top-K future FX rates instead of forecasting all the future FX rates, and bases the hold-versus-sell decision on the forecasts (e.g. hold if future FX rate is higher than current FX rate, sell otherwise). Furthermore, to handle the non-stationarity in the FX rates data which poses challenges to the i.i.d. assumption in supervised learning methods, we propose to adaptively learn decision-thresholds based on recent historical episodes. Through extensive empirical evaluation, we show that our approach is the only approach which is able to consistently improve upon a simple heuristic baseline. Further experiments show the inefficacy of state-of-the-art statistical and deep-learning-based forecasting methods as they degrade the performance of the trading agent.

Jyoti Narwariya, Chetan Verma, Pankaj Malhotra et al.

In electricity markets, retailers or brokers want to maximize profits by allocating tariff profiles to end consumers. One of the objectives of such demand response management is to incentivize the consumers to adjust their consumption so that the overall electricity procurement in the wholesale markets is minimized, e.g. it is desirable that consumers consume less during peak hours when cost of procurement for brokers from wholesale markets are high. We consider a greedy solution to maximize the overall profit for brokers by optimal tariff profile allocation. This in-turn requires forecasting electricity consumption for each user for all tariff profiles. This forecasting problem is challenging compared to standard forecasting problems due to following reasons: i. the number of possible combinations of hourly tariffs is high and retailers may not have considered all combinations in the past resulting in a biased set of tariff profiles tried in the past, ii. the profiles allocated in the past to each user is typically based on certain policy. These reasons violate the standard i.i.d. assumptions, as there is a need to evaluate new tariff profiles on existing customers and historical data is biased by the policies used in the past for tariff allocation. In this work, we consider several scenarios for forecasting and optimization under these conditions. We leverage the underlying structure of how consumers respond to variable tariff rates by comparing tariffs across hours and shifting loads, and propose suitable inductive biases in the design of deep neural network based architectures for forecasting under such scenarios. More specifically, we leverage attention mechanisms and permutation equivariant networks that allow desirable processing of tariff profiles to learn tariff representations that are insensitive to the biases in the data and still representative of the task.

Diksha Garg, Priyanka Gupta, Pankaj Malhotra et al.

Most of the existing deep reinforcement learning (RL) approaches for session-based recommendations either rely on costly online interactions with real users, or rely on potentially biased rule-based or data-driven user-behavior models for learning. In this work, we instead focus on learning recommendation policies in the pure batch or offline setting, i.e. learning policies solely from offline historical interaction logs or batch data generated from an unknown and sub-optimal behavior policy, without further access to data from the real-world or user-behavior models. We propose BCD4Rec: Batch-Constrained Distributional RL for Session-based Recommendations. BCD4Rec builds upon the recent advances in batch (offline) RL and distributional RL to learn from offline logs while dealing with the intrinsically stochastic nature of rewards from the users due to varied latent interest preferences (environments). We demonstrate that BCD4Rec significantly improves upon the behavior policy as well as strong RL and non-RL baselines in the batch setting in terms of standard performance metrics like Click Through Rates or Buy Rates. Other useful properties of BCD4Rec include: i. recommending items from the correct latent categories indicating better value estimates despite large action space (of the order of number of items), and ii. overcoming popularity bias in clicked or bought items typically present in the offline logs.

Vibhor Gupta, Jyoti Narwariya, Pankaj Malhotra et al.

Several applications of Internet of Things (IoT) technology involve capturing data from multiple sensors resulting in multi-sensor time series. Existing neural networks based approaches for such multi-sensor or multivariate time series modeling assume fixed input dimension or number of sensors. Such approaches can struggle in the practical setting where different instances of the same device or equipment such as mobiles, wearables, engines, etc. come with different combinations of installed sensors. We consider training neural network models from such multi-sensor time series, where the time series have varying input dimensionality owing to availability or installation of a different subset of sensors at each source of time series. We propose a novel neural network architecture suitable for zero-shot transfer learning allowing robust inference for multivariate time series with previously unseen combination of available dimensions or sensors at test time. Such a combinatorial generalization is achieved by conditioning the layers of a core neural network-based time series model with a "conditioning vector" that carries information of the available combination of sensors for each time series. This conditioning vector is obtained by summarizing the set of learned "sensor embedding vectors" corresponding to the available sensors in a time series via a graph neural network. We evaluate the proposed approach on publicly available activity recognition and equipment prognostics datasets, and show that the proposed approach allows for better generalization in comparison to a deep gated recurrent neural network baseline.

Jyoti Narwariya, Pankaj Malhotra, Vishnu TV et al.

Automated equipment health monitoring from streaming multisensor time-series data can be used to enable condition-based maintenance, avoid sudden catastrophic failures, and ensure high operational availability. We note that most complex machinery has a well-documented and readily accessible underlying structure capturing the inter-dependencies between sub-systems or modules. Deep learning models such as those based on recurrent neural networks (RNNs) or convolutional neural networks (CNNs) fail to explicitly leverage this potentially rich source of domain-knowledge into the learning procedure. In this work, we propose to capture the structure of a complex equipment in the form of a graph, and use graph neural networks (GNNs) to model multi-sensor time-series data. Using remaining useful life estimation as an application task, we evaluate the advantage of incorporating the graph structure via GNNs on the publicly available turbofan engine benchmark dataset. We observe that the proposed GNN-based RUL estimation model compares favorably to several strong baselines from literature such as those based on RNNs and CNNs. Additionally, we observe that the learned network is able to focus on the module (node) with impending failure through a simple attention mechanism, potentially paving the way for actionable diagnosis.

Jyoti Narwariya, Pankaj Malhotra, Lovekesh Vig et al.

Deep neural networks (DNNs) have achieved state-of-the-art results on time series classification (TSC) tasks. In this work, we focus on leveraging DNNs in the often-encountered practical scenario where access to labeled training data is difficult, and where DNNs would be prone to overfitting. We leverage recent advancements in gradient-based meta-learning, and propose an approach to train a residual neural network with convolutional layers as a meta-learning agent for few-shot TSC. The network is trained on a diverse set of few-shot tasks sampled from various domains (e.g. healthcare, activity recognition, etc.) such that it can solve a target task from another domain using only a small number of training samples from the target task. Most existing meta-learning approaches are limited in practice as they assume a fixed number of target classes across tasks. We overcome this limitation in order to train a common agent across domains with each domain having different number of target classes, we utilize a triplet-loss based learning procedure that does not require any constraints to be enforced on the number of classes for the few-shot TSC tasks. To the best of our knowledge, we are the first to use meta-learning based pre-training for TSC. Our approach sets a new benchmark for few-shot TSC, outperforming several strong baselines on few-shot tasks sampled from 41 datasets in UCR TSC Archive. We observe that pre-training under the meta-learning paradigm allows the network to quickly adapt to new unseen tasks with small number of labeled instances.

Vishnu TV, Pankaj Malhotra, Jyoti Narwariya et al.

Recently, neural networks trained as optimizers under the "learning to learn" or meta-learning framework have been shown to be effective for a broad range of optimization tasks including derivative-free black-box function optimization. Recurrent neural networks (RNNs) trained to optimize a diverse set of synthetic non-convex differentiable functions via gradient descent have been effective at optimizing derivative-free black-box functions. In this work, we propose RNN-Opt: an approach for learning RNN-based optimizers for optimizing real-parameter single-objective continuous functions under limited budget constraints. Existing approaches utilize an observed improvement based meta-learning loss function for training such models. We propose training RNN-Opt by using synthetic non-convex functions with known (approximate) optimal values by directly using discounted regret as our meta-learning loss function. We hypothesize that a regret-based loss function mimics typical testing scenarios, and would therefore lead to better optimizers compared to optimizers trained only to propose queries that improve over previous queries. Further, RNN-Opt incorporates simple yet effective enhancements during training and inference procedures to deal with the following practical challenges: i) Unknown range of possible values for the black-box function to be optimized, and ii) Practical and domain-knowledge based constraints on the input parameters. We demonstrate the efficacy of RNN-Opt in comparison to existing methods on several synthetic as well as standard benchmark black-box functions along with an anonymized industrial constrained optimization problem.

Kathan Kashiparekh, Jyoti Narwariya, Pankaj Malhotra et al.

Training deep neural networks often requires careful hyper-parameter tuning and significant computational resources. In this paper, we propose ConvTimeNet (CTN): an off-the-shelf deep convolutional neural network (CNN) trained on diverse univariate time series classification (TSC) source tasks. Once trained, CTN can be easily adapted to new TSC target tasks via a small amount of fine-tuning using labeled instances from the target tasks. We note that the length of convolutional filters is a key aspect when building a pre-trained model that can generalize to time series of different lengths across datasets. To achieve this, we incorporate filters of multiple lengths in all convolutional layers of CTN to capture temporal features at multiple time scales. We consider all 65 datasets with time series of lengths up to 512 points from the UCR TSC Benchmark for training and testing transferability of CTN: We train CTN on a randomly chosen subset of 24 datasets using a multi-head approach with a different softmax layer for each training dataset, and study generalizability and transferability of the learned filters on the remaining 41 TSC datasets. We observe significant gains in classification accuracy as well as computational efficiency when using pre-trained CTN as a starting point for subsequent task-specific fine-tuning compared to existing state-of-the-art TSC approaches. We also provide qualitative insights into the working of CTN by: i) analyzing the activations and filters of first convolution layer suggesting the filters in CTN are generically useful, ii) analyzing the impact of the design decision to incorporate multiple length decisions, and iii) finding regions of time series that affect the final classification decision via occlusion sensitivity analysis.

Priyanka Gupta, Pankaj Malhotra, Jyoti Narwariya et al.

Deep neural networks have shown promising results for various clinical prediction tasks. However, training deep networks such as those based on Recurrent Neural Networks (RNNs) requires large labeled data, significant hyper-parameter tuning effort and expertise, and high computational resources. In this work, we investigate as to what extent can transfer learning address these issues when using deep RNNs to model multivariate clinical time series. We consider two scenarios for transfer learning using RNNs: i) domain-adaptation, i.e., leveraging a deep RNN - namely, TimeNet - pre-trained for feature extraction on time series from diverse domains, and adapting it for feature extraction and subsequent target tasks in healthcare domain, ii) task-adaptation, i.e., pre-training a deep RNN - namely, HealthNet - on diverse tasks in healthcare domain, and adapting it to new target tasks in the same domain. We evaluate the above approaches on publicly available MIMIC-III benchmark dataset, and demonstrate that (a) computationally-efficient linear models trained using features extracted via pre-trained RNNs outperform or, in the worst case, perform as well as deep RNNs and statistical hand-crafted features based models trained specifically for target task; (b) models obtained by adapting pre-trained models for target tasks are significantly more robust to the size of labeled data compared to task-specific RNNs, while also being computationally efficient. We, therefore, conclude that pre-trained deep models like TimeNet and HealthNet allow leveraging the advantages of deep learning for clinical time series analysis tasks, while also minimize dependence on hand-crafted features, deal robustly with scarce labeled training data scenarios without overfitting, as well as reduce dependence on expertise and resources required to train deep networks from scratch.

Vishnu TV, Diksha, Pankaj Malhotra et al.

Prognostics or Remaining Useful Life (RUL) Estimation from multi-sensor time series data is useful to enable condition-based maintenance and ensure high operational availability of equipment. We propose a novel deep learning based approach for Prognostics with Uncertainty Quantification that is useful in scenarios where: (i) access to labeled failure data is scarce due to rarity of failures (ii) future operational conditions are unobserved and (iii) inherent noise is present in the sensor readings. All three scenarios mentioned are unavoidable sources of uncertainty in the RUL estimation process often resulting in unreliable RUL estimates. To address (i), we formulate RUL estimation as an Ordinal Regression (OR) problem, and propose LSTM-OR: deep Long Short Term Memory (LSTM) network based approach to learn the OR function. We show that LSTM-OR naturally allows for incorporation of censored operational instances in training along with the failed instances, leading to more robust learning. To address (ii), we propose a simple yet effective approach to quantify predictive uncertainty in the RUL estimation models by training an ensemble of LSTM-OR models. Through empirical evaluation on C-MAPSS turbofan engine benchmark datasets, we demonstrate that LSTM-OR is significantly better than the commonly used deep metric regression based approaches for RUL estimation, especially when failed training instances are scarce. Further, our uncertainty quantification approach yields high quality predictive uncertainty estimates while also leading to improved RUL estimates compared to single best LSTM-OR models.

Priyanka Gupta, Pankaj Malhotra, Lovekesh Vig et al.

Deep neural networks have shown promising results for various clinical prediction tasks such as diagnosis, mortality prediction, predicting duration of stay in hospital, etc. However, training deep networks -- such as those based on Recurrent Neural Networks (RNNs) -- requires large labeled data, high computational resources, and significant hyperparameter tuning effort. In this work, we investigate as to what extent can transfer learning address these issues when using deep RNNs to model multivariate clinical time series. We consider transferring the knowledge captured in an RNN trained on several source tasks simultaneously using a large labeled dataset to build the model for a target task with limited labeled data. An RNN pre-trained on several tasks provides generic features, which are then used to build simpler linear models for new target tasks without training task-specific RNNs. For evaluation, we train a deep RNN to identify several patient phenotypes on time series from MIMIC-III database, and then use the features extracted using that RNN to build classifiers for identifying previously unseen phenotypes, and also for a seemingly unrelated task of in-hospital mortality. We demonstrate that (i) models trained on features extracted using pre-trained RNN outperform or, in the worst case, perform as well as task-specific RNNs; (ii) the models using features from pre-trained models are more robust to the size of labeled data than task-specific RNNs; and (iii) features extracted using pre-trained RNN are generic enough and perform better than typical statistical hand-crafted features.

Narendhar Gugulothu, Vishnu TV, Pankaj Malhotra et al.

We consider the problem of estimating the remaining useful life (RUL) of a system or a machine from sensor data. Many approaches for RUL estimation based on sensor data make assumptions about how machines degrade. Additionally, sensor data from machines is noisy and often suffers from missing values in many practical settings. We propose Embed-RUL: a novel approach for RUL estimation from sensor data that does not rely on any degradation-trend assumptions, is robust to noise, and handles missing values. Embed-RUL utilizes a sequence-to-sequence model based on Recurrent Neural Networks (RNNs) to generate embeddings for multivariate time series subsequences. The embeddings for normal and degraded machines tend to be different, and are therefore found to be useful for RUL estimation. We show that the embeddings capture the overall pattern in the time series while filtering out the noise, so that the embeddings of two machines with similar operational behavior are close to each other, even when their sensor readings have significant and varying levels of noise content. We perform experiments on publicly available turbofan engine dataset and a proprietary real-world dataset, and demonstrate that Embed-RUL outperforms the previously reported state-of-the-art on several metrics.

Pankaj Malhotra, Vishnu TV, Lovekesh Vig et al.

Inspired by the tremendous success of deep Convolutional Neural Networks as generic feature extractors for images, we propose TimeNet: a deep recurrent neural network (RNN) trained on diverse time series in an unsupervised manner using sequence to sequence (seq2seq) models to extract features from time series. Rather than relying on data from the problem domain, TimeNet attempts to generalize time series representation across domains by ingesting time series from several domains simultaneously. Once trained, TimeNet can be used as a generic off-the-shelf feature extractor for time series. The representations or embeddings given by a pre-trained TimeNet are found to be useful for time series classification (TSC). For several publicly available datasets from UCR TSC Archive and an industrial telematics sensor data from vehicles, we observe that a classifier learned over the TimeNet embeddings yields significantly better performance compared to (i) a classifier learned over the embeddings given by a domain-specific RNN, as well as (ii) a nearest neighbor classifier based on Dynamic Time Warping.

Pankaj Malhotra, Vishnu TV, Anusha Ramakrishnan et al.

Many approaches for estimation of Remaining Useful Life (RUL) of a machine, using its operational sensor data, make assumptions about how a system degrades or a fault evolves, e.g., exponential degradation. However, in many domains degradation may not follow a pattern. We propose a Long Short Term Memory based Encoder-Decoder (LSTM-ED) scheme to obtain an unsupervised health index (HI) for a system using multi-sensor time-series data. LSTM-ED is trained to reconstruct the time-series corresponding to healthy state of a system. The reconstruction error is used to compute HI which is then used for RUL estimation. We evaluate our approach on publicly available Turbofan Engine and Milling Machine datasets. We also present results on a real-world industry dataset from a pulverizer mill where we find significant correlation between LSTM-ED based HI and maintenance costs.

Pankaj Malhotra, Puneet Agarwal, Gautam Shroff

In many government applications we often find that information about entities, such as persons, are available in disparate data sources such as passports, driving licences, bank accounts, and income tax records. Similar scenarios are commonplace in large enterprises having multiple customer, supplier, or partner databases. Each data source maintains different aspects of an entity, and resolving entities based on these attributes is a well-studied problem. However, in many cases documents in one source reference those in others; e.g., a person may provide his driving-licence number while applying for a passport, or vice-versa. These links define relationships between documents of the same entity (as opposed to inter-entity relationships, which are also often used for resolution). In this paper we describe an algorithm to cluster documents that are highly likely to belong to the same entity by exploiting inter-document references in addition to attribute similarity. Our technique uses a combination of iterative graph-traversal, locality-sensitive hashing, iterative match-merge, and graph-clustering to discover unique entities based on a document corpus. A unique feature of our technique is that new sets of documents can be added incrementally while having to re-resolve only a small subset of a previously resolved entity-document collection. We present performance and quality results on two data-sets: a real-world database of companies and a large synthetically generated `population' database. We also demonstrate benefit of using inter-document references for clustering in the form of enhanced recall of documents for resolution.