Search

We model short-duration (e.g. day) trading in financial markets as a sequential decision-making problem under uncertainty, with the added complication of continual concept-drift. We, therefore, employ meta reinforcement learning via the RL2 algorithm. It is also known that human traders often rely on frequently occurring symbolic patterns in price series. We employ logical program induction to discover symbolic patterns that occur frequently as well as recently, and explore whether using such features improves the performance of our meta reinforcement learning algorithm. We report experiments on real data indicating that meta-RL is better than vanilla RL and also benefits from learned symbolic features., Comment: To appear in Muffin@AAAI'23

When answering natural language questions over knowledge bases, missing facts, incomplete schema and limited scope naturally lead to many questions being unanswerable. While answerability has been explored in other QA settings, it has not been studied for QA over knowledge bases (KBQA). We create GrailQAbility, a new benchmark KBQA dataset with unanswerability, by first identifying various forms of KB incompleteness that make questions unanswerable, and then systematically adapting GrailQA (a popular KBQA dataset with only answerable questions). Experimenting with three state-of-the-art KBQA models, we find that all three models suffer a drop in performance even after suitable adaptation for unanswerable questions. In addition, these often detect unanswerability for wrong reasons and find specific forms of unanswerability particularly difficult to handle. This underscores the need for further research in making KBQA systems robust to unanswerability

We demonstrate a Physics-informed Neural Network (PINN) based model for real-time health monitoring of a heat exchanger, that plays a critical role in improving energy efficiency of thermal power plants. A hypernetwork based approach is used to enable the domain-decomposed PINN learn the thermal behavior of the heat exchanger in response to dynamic boundary conditions, eliminating the need to re-train. As a result, we achieve orders of magnitude reduction in inference time in comparison to existing PINNs, while maintaining the accuracy on par with the physics-based simulations. This makes the approach very attractive for predictive maintenance of the heat exchanger in digital twin environments., Comment: Neural Information Processing Systems 2022: The Machine Learning and the Physical Sciences workshop

Deep neural networks (DNN) are prone to miscalibrated predictions, often exhibiting a mismatch between the predicted output and the associated confidence scores. Contemporary model calibration techniques mitigate the problem of overconfident predictions by pushing down the confidence of the winning class while increasing the confidence of the remaining classes across all test samples. However, from a deployment perspective, an ideal model is desired to (i) generate well-calibrated predictions for high-confidence samples with predicted probability say >0.95, and (ii) generate a higher proportion of legitimate high-confidence samples. To this end, we propose a novel regularization technique that can be used with classification losses, leading to state-of-the-art calibrated predictions at test time; From a deployment standpoint in safety-critical applications, only high-confidence samples from a well-calibrated model are of interest, as the remaining samples have to undergo manual inspection. Predictive confidence reduction of these potentially ``high-confidence samples'' is a downside of existing calibration approaches. We mitigate this by proposing a dynamic train-time data pruning strategy that prunes low-confidence samples every few epochs, providing an increase in "confident yet calibrated samples". We demonstrate state-of-the-art calibration performance across image classification benchmarks, reducing training time without much compromise in accuracy. We provide insights into why our dynamic pruning strategy that prunes low-confidence training samples leads to an increase in high-confidence samples at test time., Comment: The paper is accepted at Winter Conference on applications of Computer Vision (IEEE WACV) in algorithms tracks. 8 pages Main paper; 3 pages supplementary material

The first order transition between the confining and the center symmetry breaking phases of the SU(3) Yang-Mills theory is marked by discontinuities in various thermodynamics functions, such as the energy density or the value of the Polyakov loop. We investigate the non-analytical behaviour of the topological susceptibility and its higher cumulant around the transition temperature and make the connection to the curvature of the phase diagram in the $T-\theta$ plane and to the latent heat., Comment: 13 pages, 10 figures

We are interested in neurosymbolic systems consisting of a high-level symbolic layer for explainable prediction in terms of human-intelligible concepts; and a low-level neural layer for extracting symbols required to generate the symbolic explanation. Real data is often imperfect meaning that even if the symbolic theory remains unchanged, we may still need to address the problem of mapping raw data to high-level symbols, each time there is a change in the data acquisition environment or equipment. Manual (re-)annotation of the raw data each time this happens is laborious and expensive; and automated labelling methods are often imperfect, especially for complex problems. NEUROLOG proposed the use of a semantic loss function that allows an existing feature-based symbolic model to guide the extraction of feature-values from raw data, using `abduction'. However, the experiments demonstrating the use of semantic loss through abduction appear to rely heavily on a domain-specific pre-processing step that enables a prior delineation of feature locations in the raw data. We examine the use of semantic loss in domains where such pre-processing is not possible, or is not obvious. We show that without any prior information about the features, the NEUROLOG approach can continue to predict accurately even with substantially incorrect feature predictions. We show also that prior information about the features in the form of even imperfect pre-training can help correct this situation. These findings are replicated on the original problem considered by NEUROLOG, without the use of feature-delineation. This suggests that symbolic explanations constructed for data in a domain could be re-used in a related domain, by `feature-adaptation' of pre-trained neural extractors using the semantic loss function constrained by abductive feedback.

The HH80-81 system is one of the most powerful jets driven by a massive protostar. We present new near-infrared (NIR) line imaging observations of the HH80-81 jet in the H$_2$ (2.122 $\mu$m) and [Fe II] (1.644 $\mu$m) lines. These lines trace not only the jet close to the exciting source but also the knots located farther away. We have detected nine groups of knot-like structures in the jet including HH80 and HH81 spaced $0.2-0.9$ pc apart. The knots in the northern arm of the jet show only [Fe II] emission closer to the exciting source, a combination of [Fe II] and H$_2$ at intermediate distances, and solely H$_2$ emission farther outwards. Towards the southern arm, all the knots exhibit both H$_2$ and [Fe II] emission. The nature of the shocks is inferred by assimilating the NIR observations with radio and X-ray observations from literature. In the northern arm, we infer the presence of strong dissociative shocks, in the knots located close to the exciting source. The knots in the southern arm that include HH80 and HH81 are explicable as a combination of strong and weak shocks. The mass-loss rates of the knots determined from [Fe II] luminosities are in the range $\sim 3.0\times 10^{-7}-5.2\times 10^{-5}$ M$_{\odot}$ yr$^{-1}$, consistent with those from massive protostars. Towards the central region, close to the driving source of the jet, we have observed various arcs in H$_2$ emission which resemble bow shocks, and strings of H$_2$ knots which reveal traces of multiple outflows., Comment: 20 pages, 5 figures, 3 tables, Accepted for publication in The Astrophysical Journal

State-of-the-art summarization models still struggle to be factually consistent with the input text. A model-agnostic way to address this problem is post-editing the generated summaries. However, existing approaches typically fail to remove entity errors if a suitable input entity replacement is not available or may insert erroneous content. In our work, we focus on removing extrinsic entity errors, or entities not in the source, to improve consistency while retaining the summary's essential information and form. We propose to use sentence-compression data to train the post-editing model to take a summary with extrinsic entity errors marked with special tokens and output a compressed, well-formed summary with those errors removed. We show that this model improves factual consistency while maintaining ROUGE, improving entity precision by up to 30% on XSum, and that this model can be applied on top of another post-editor, improving entity precision by up to a total of 38%. We perform an extensive comparison of post-editing approaches that demonstrate trade-offs between factual consistency, informativeness, and grammaticality, and we analyze settings where post-editors show the largest improvements., Comment: EMNLP 2022

Analogical Reasoning problems challenge both connectionist and symbolic AI systems as these entail a combination of background knowledge, reasoning and pattern recognition. While symbolic systems ingest explicit domain knowledge and perform deductive reasoning, they are sensitive to noise and require inputs be mapped to preset symbolic features. Connectionist systems on the other hand can directly ingest rich input spaces such as images, text or speech and recognize pattern even with noisy inputs. However, connectionist models struggle to include explicit domain knowledge for deductive reasoning. In this paper, we propose a framework that combines the pattern recognition abilities of neural networks with symbolic reasoning and background knowledge for solving a class of Analogical Reasoning problems where the set of attributes and possible relations across them are known apriori. We take inspiration from the 'neural algorithmic reasoning' approach [DeepMind 2020] and use problem-specific background knowledge by (i) learning a distributed representation based on a symbolic model of the problem (ii) training neural-network transformations reflective of the relations involved in the problem and finally (iii) training a neural network encoder from images to the distributed representation in (i). These three elements enable us to perform search-based reasoning using neural networks as elementary functions manipulating distributed representations. We test this on visual analogy problems in RAVENs Progressive Matrices, and achieve accuracy competitive with human performance and, in certain cases, superior to initial end-to-end neural-network based approaches. While recent neural models trained at scale yield SOTA, our novel neuro-symbolic reasoning approach is a promising direction for this problem, and is arguably more general, especially for problems where domain knowledge is available., Comment: 13 pages, 4 figures, Accepted at 16th International Workshop on Neural-Symbolic Learning and Reasoning as part of the 2nd International Joint Conference on Learning & Reasoning (IJCLR 2022)

The star-forming activity in the HII region RCW 42 is investigated using multiple wavebands, from near-infrared to radio wavelengths. Located at a distance of 5.8 kpc, this southern region has a bolometric luminosity of 1.8 $\times$ 10$^6$ L$_{\odot}$. The ionized gas emission has been imaged at low radio frequencies of 610 and 1280 MHz using the Giant Metrewave Radio Telescope, India and shows a large expanse of the HII region, spanning $20\times 15$ pc$^2$. The average electron number density in the region is estimated to be $\sim70$ cm$^{-3}$, which suggests an average ionization fraction of the cloud to be $11\%$. An extended green object EGO G274.0649-01.1460 and several young stellar objects have been identified in the region using data from the 2MASS and Spitzer surveys. The dust emission from the associated molecular cloud is probed using Herschel Space Telescope, which reveals the presence of 5 clumps, C1-C5, in this region. Two millimetre emission cores of masses 380 and 390 M$_{\odot}$ towards the radio emission peak have been identified towards C1 from the ALMA map at 1.4 mm. The clumps are investigated for their evolutionary stages based on association with various star-formation tracers, and we find that all the clumps are in active/evolved stage., Comment: 14 pages, 13 figures, 3 tables, Accepted by MNRAS

Fasteners play a critical role in securing various parts of machinery. Deformations such as dents, cracks, and scratches on the surface of fasteners are caused by material properties and incorrect handling of equipment during production processes. As a result, quality control is required to ensure safe and reliable operations. The existing defect inspection method relies on manual examination, which consumes a significant amount of time, money, and other resources; also, accuracy cannot be guaranteed due to human error. Automatic defect detection systems have proven impactful over the manual inspection technique for defect analysis. However, computational techniques such as convolutional neural networks (CNN) and deep learning-based approaches are evolutionary methods. By carefully selecting the design parameter values, the full potential of CNN can be realised. Using Taguchi-based design of experiments and analysis, an attempt has been made to develop a robust automatic system in this study. The dataset used to train the system has been created manually for M14 size nuts having two labeled classes: Defective and Non-defective. There are a total of 264 images in the dataset. The proposed sequential CNN comes up with a 96.3% validation accuracy, 0.277 validation loss at 0.001 learning rate., Comment: 13 pages, 6 figures

We introduce Physics Informed Symbolic Networks (PISN) which utilize physics-informed loss to obtain a symbolic solution for a system of Partial Differential Equations (PDE). Given a context-free grammar to describe the language of symbolic expressions, we propose to use weighted sum as continuous approximation for selection of a production rule. We use this approximation to define multilayer symbolic networks. We consider Kovasznay flow (Navier-Stokes) and two-dimensional viscous Burger's equations to illustrate that PISN are able to provide a performance comparable to PINNs across various start-of-the-art advances: multiple outputs and governing equations, domain-decomposition, hypernetworks. Furthermore, we propose Physics-informed Neurosymbolic Networks (PINSN) which employ a multilayer perceptron (MLP) operator to model the residue of symbolic networks. PINSNs are observed to give 2-3 orders of performance gain over standard PINN., Comment: Neural Information Processing Systems 2022: The Symbiosis of Deep Learning and Differential Equations Workshop

Globular clusters being old and densely packed serve as ideal laboratories to test stellar evolution theories. Although there is enormous literature on globular clusters in optical bands, studies in the ultraviolet (UV) regime are sparse. In this work, we study the stellar populations of a metal poor and a rather dispersed globular cluster, NGC 5053, using the UV instrument of AstroSat, namely the Ultra Violet Imaging Telescope in three far-UV (F154W, F169M, F172M) and three near-UV (N219M, N245M, N263M) filters. Photometry was carried out on these images to construct a catalogue of UV stars, of which the cluster members were identified using Gaia EDR3 catalogue. UV and optical CMDs help us locate known stellar populations such as BHB stars, RR-Lyrae stars, RHB stars, BSSs, SX-Phe, RGB and AGB stars. Based on their locations in the CMDs, we have identified 8 new BSS candidates, 6 probable eBSSs, and an EHB candidate. Their nature has been confirmed by fitting their spectral energy distributions with stellar atmospheric models. We believe the BSS population of this cluster is likely to have a collisional origin based on our analyses of their radial distribution and SEDs. BaSTI-IAC isochrones were generated to characterize the cluster properties, and we find that the observed brightness and colours of cluster members are best-fit with a model that is alpha-enhanced with a helium fraction of 0.247, metallicity of -1.9 dex and age within a range of 10.5-14.5 Gyr., Comment: Accepted for publication in MNRAS, 14 pages, 12 figures, 3 Tables

Deep learning models for natural language processing (NLP) are increasingly adopted and deployed by analysts without formal training in NLP or machine learning (ML). However, the documentation intended to convey the model's details and appropriate use is tailored primarily to individuals with ML or NLP expertise. To address this gap, we conduct a design inquiry into interactive model cards, which augment traditionally static model cards with affordances for exploring model documentation and interacting with the models themselves. Our investigation consists of an initial conceptual study with experts in ML, NLP, and AI Ethics, followed by a separate evaluative study with non-expert analysts who use ML models in their work. Using a semi-structured interview format coupled with a think-aloud protocol, we collected feedback from a total of 30 participants who engaged with different versions of standard and interactive model cards. Through a thematic analysis of the collected data, we identified several conceptual dimensions that summarize the strengths and limitations of standard and interactive model cards, including: stakeholders; design; guidance; understandability & interpretability; sensemaking & skepticism; and trust & safety. Our findings demonstrate the importance of carefully considered design and interactivity for orienting and supporting non-expert analysts using deep learning models, along with a need for consideration of broader sociotechnical contexts and organizational dynamics. We have also identified design elements, such as language, visual cues, and warnings, among others, that support interactivity and make non-interactive content accessible. We summarize our findings as design guidelines and discuss their implications for a human-centered approach towards AI/ML documentation., Comment: To appear at ACM FAccT'22

Protostellar jets and outflows are pointers of star-formation and serve as important sources of momentum and energy transfer to the interstellar medium. Radio emission from ionized jets have been detected towards a number of protostellar objects. In few cases, negative spectral indices and polarized emission have also been observed suggesting the presence of synchrotron emission from relativistic electrons. In this work, we develop a numerical model that incorporates both thermal free-free and non-thermal synchrotron emission mechanisms in the jet geometry. The flux densities include contribution from an inner thermal jet, and a combination of emission from thermal and non-thermal distributions along the edges and extremities, where the jet interacts with the interstellar medium. We also include the effect of varying ionization fraction laterally across the jet. An investigation of radio emission and spectra along the jet shows the dependence of the emission process and optical depth along the line of sight. We explore the effect of various parameters on the turnover frequencies and the radio spectral indices (between 10 MHz and 300 GHz) associated with them., Comment: 18 pages, 14 figures, 2 Tables. Accepted for publication in MNRAS

With over 50 million car sales annually and over 1.3 million deaths every year due to motor accidents we have chosen this space. India accounts for 11 per cent of global death in road accidents. Drivers are held responsible for 78% of accidents. Road safety problems in developing countries is a major concern and human behavior is ascribed as one of the main causes and accelerators of road safety problems. Driver distraction has been identified as the main reason for accidents. Distractions can be caused due to reasons such as mobile usage, drinking, operating instruments, facial makeup, social interaction. For the scope of this project, we will focus on building a highly efficient ML model to classify different driver distractions at runtime using computer vision. We would also analyze the overall speed and scalability of the model in order to be able to set it up on an edge device. We use CNN, VGG-16, RestNet50 and ensemble of CNN to predict the classes.

Existing universal lesion detection (ULD) methods utilize compute-intensive anchor-based architectures which rely on predefined anchor boxes, resulting in unsatisfactory detection performance, especially in small and mid-sized lesions. Further, these default fixed anchor-sizes and ratios do not generalize well to different datasets. Therefore, we propose a robust one-stage anchor-free lesion detection network that can perform well across varying lesions sizes by exploiting the fact that the box predictions can be sorted for relevance based on their center rather than their overlap with the object. Furthermore, we demonstrate that the ULD can be improved by explicitly providing it the domain-specific information in the form of multi-intensity images generated using multiple HU windows, followed by self-attention based feature-fusion and backbone initialization using weights learned via self-supervision over CT-scans. We obtain comparable results to the state-of-the-art methods, achieving an overall sensitivity of 86.05% on the DeepLesion dataset, which comprises of approximately 32K CT-scans with lesions annotated across various body organs., Comment: 4 Pages, 2 figures, 2 tables. Paper accepted at IEEE International Symposium on Biomedical Imaging (ISBI'22)

Incorporating data-specific domain knowledge in deep networks explicitly can provide important cues beneficial for lesion detection and can mitigate the need for diverse heterogeneous datasets for learning robust detectors. In this paper, we exploit the domain information present in computed tomography (CT) scans and propose a robust universal lesion detection (ULD) network that can detect lesions across all organs of the body by training on a single dataset, DeepLesion. We analyze CT-slices of varying intensities, generated using heuristically determined Hounsfield Unit(HU) windows that individually highlight different organs and are given as inputs to the deep network. The features obtained from the multiple intensity images are fused using a novel convolution augmented multi-head self-attention module and subsequently, passed to a Region Proposal Network (RPN) for lesion detection. In addition, we observed that traditional anchor boxes used in RPN for natural images are not suitable for lesion sizes often found in medical images. Therefore, we propose to use lesion-specific anchor sizes and ratios in the RPN for improving the detection performance. We use self-supervision to initialize weights of our network on the DeepLesion dataset to further imbibe domain knowledge. Our proposed Domain Knowledge augmented Multi-head Attention based Universal Lesion Detection Network DMKA-ULD produces refined and precise bounding boxes around lesions across different organs. We evaluate the efficacy of our network on the publicly available DeepLesion dataset which comprises of approximately 32K CT scans with annotated lesions across all organs of the body. Results demonstrate that we outperform existing state-of-the-art methods achieving an overall sensitivity of 87.16%., Comment: Main Paper: 13 Pages, 5 Figures, 2 Tables. Supplementary: 4 Pages, 1 Figure, 3 Tables. Paper accepted at The 32nd British Machine Vision Conference (BMVC'21)