Papers by Dinesh Khandelwal
Neurology India
Rasmussen's encephalopathy (RE) is an uncommon neurological disease of inflammatory origin wh... more Rasmussen's encephalopathy (RE) is an uncommon neurological disease of inflammatory origin which is characterized by intractable focal epilepsy, progressive limb weakness, and cognitive deterioration. RE presenting as movement disorder like hemidystonia or hemichorea is a rare occurrence. The duration of prodromal stage of RE is usually in weeks or months. Prolonged prodromal stage like in years is rarely reported. Magnetic resonance imaging (MRI) is a good biomarker in RE and it also suggests the sequential progression of disease. Here we report two cases of RE, one presenting with hemidystonia and other case with unusually prolonged prodromal stage duration of 7 years. In spite of severe hemi-atrophy of brain in second case response to immunomodulators was dramatic.
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)
ArXiv, 2020
Knowledge base question answering (KBQA) is an important task in Natural Language Processing. Exi... more Knowledge base question answering (KBQA) is an important task in Natural Language Processing. Existing approaches face significant challenges including complex question understanding, necessity for reasoning, and lack of large training datasets. In this work, we propose a semantic parsing and reasoning-based Neuro-Symbolic Question Answering(NSQA) system, that leverages (1) Abstract Meaning Representation (AMR) parses for task-independent question under-standing; (2) a novel path-based approach to transform AMR parses into candidate logical queries that are aligned to the KB; (3) a neuro-symbolic reasoner called Logical Neural Net-work (LNN) that executes logical queries and reasons over KB facts to provide an answer; (4) system of systems approach,which integrates multiple, reusable modules that are trained specifically for their individual tasks (e.g. semantic parsing,entity linking, and relationship linking) and do not require end-to-end training data. NSQA achieves state-of-the-...
The Egyptian Journal of Neurology, Psychiatry and Neurosurgery
Background Cerebral arterial thromboses or ischemic strokes may be caused by cumulative or indepe... more Background Cerebral arterial thromboses or ischemic strokes may be caused by cumulative or independent effects of a variety of risk factors. High factor VIII level is one of those important but less known risk factors for arterial and venous thrombosis. We hereby provide a comprehensive review of the role of high factor VIII levels as a risk factor of arterial thrombosis. Moreover, we present our views on inclusion of factor VIII testing in the etiology workup protocol of young patients with ischemic strokes and their treatment with anticoagulant therapy. Case presentation We illustrate a case of 32-year-old North Indian female patient with Ischemic stroke whose only identifiable risk factor was revealed to be an elevated factor VIII level. She was treated with oral anticoagulant with an uneventful follow-up of 6 months. Conclusions Elevated factor VIII levels have their independent and additive effects in causation and prognosis of arterial strokes. We herein discuss the mechanism ...
Quantum logic inspired embedding (aka Quantum Embedding (QE)) of a Knowledge-Base (KB) was propos... more Quantum logic inspired embedding (aka Quantum Embedding (QE)) of a Knowledge-Base (KB) was proposed recently by Garg et al. [1]. It is claimed that the QE preserves the logical structure of the input KB given in the form of unary and binary predicates hierarchy. Such structure preservation allows one to perform Boolean logic style deductive reasoning directly over these embedding vectors. The original QE idea, however, is limited to the transductive (not inductive) setting. Moreover, the original QE scheme runs quite slow on real applications involving millions of entities. This paper alleviates both of these key limitations. We start by reformulating the original QE problem to allow for the induction. On the way, we also underscore some interesting analytic and geometric properties of the solution and leverage them to design a faster training scheme. As an application, we show that one can achieve state-of-the-art performance on the well-known NLP task of fine-grained entity type c...
International journal of scientific research, 2019
Inflammatory neuropathies such as Guillain-Barré syndrome (GBS) and chronic inflammatory demyelin... more Inflammatory neuropathies such as Guillain-Barré syndrome (GBS) and chronic inflammatory demyelinating polyneuropathy (CIDP) are well-known immune-mediated polyneuropathies sharing many common characteristics during the acute disease phase. Autoimmune disorders are believed to develop as a result of interplay between genetic and environmental factors. GBS is often associated with preceding infections such as Campylobacter jejuni (C. jejuni), Mycoplasma pneumoniae, Epstein-Barr virus, cytomegalovirus (CMV), and varicella zoster virus (VZV). 1 In contrast, it is challenging to find an antecedent infection as a specific triggering event for CIDP, even in acute-onset CIDP (A-CIDP). 2 This might be due to environmental factors influencing the development of GBS more than the development of CIDP, and the immune system of CIDP patients being genetically permissive to be activated and, once activated, letting autoreactive T-cells remain viable and activated to cause a chronic autoimmune disease. 3,4 Here we describe a patient who had initially been diagnosed with GBS following hepatitis A virus (HAV) infection, but whose diagnosis was eventually changed to A-CIDP.
ArXiv, 2018
Many prediction tasks, especially in computer vision, are often inherently ambiguous. For example... more Many prediction tasks, especially in computer vision, are often inherently ambiguous. For example, the output of semantic segmentation may depend on the scale one is looking at, and image saliency or video summarization is often user or context dependent. Arguably, in such scenarios, exploiting instance specific evidence, such as scale or user context, can help resolve the underlying ambiguity leading to the improved predictions. While existing literature has considered incorporating such evidence in classical models such as probabilistic graphical models (PGMs), there is limited (or no) prior work looking at this problem in the context of deep neural network (DNN) models. In this paper, we present a generic multi task learning (MTL) based framework which handles the evidence as the output of one or more secondary tasks, while modeling the original problem as the primary task of interest. Our training phase is identical to the one used by standard MTL architectures. During predictio...
A theorem involving a subspace can be interpreted as theorem about orthogonal projection matrix i... more A theorem involving a subspace can be interpreted as theorem about orthogonal projection matrix in the sense that subspace may be expressed as a linear transformation on R [1]. Therefore we formulate the IQE problem in term of orthogonal projection matrices. To formulate the IQE problem as a non-convex optimization problem, we need some facts about orthogonal projection matrices. In this section, we review some important concepts and key results about orthogonal projection matrices. We have sketched out the proofs for some of the theorems, while the proof for other theorems can be found in [2], [3], [4].
Objective: To Study the Clinical, Psychological and Neurological Profile and Assessing the Outcom... more Objective: To Study the Clinical, Psychological and Neurological Profile and Assessing the Outcome Predictors of Patients with Psychogenic Non Epileptic Seizures in North Indian adult population. Design of study: A prospective observational study, conducted at tertiary teaching institute at Jaipur. Materials and Methods: Seventy-four patients with PNES were enrolled. The diagnosis was based on mobile recordings of episodes, and video-electroencephalography (video-EEG) monitoring. A detailed clinical evaluation was done including evaluation for coexistent anxiety, depression or other psychiatric disorder. Patients were followed-up to 2 months. Statistical Analysis: Appropriate statistical tests were used. Results: The mean age at onset was 22 years; with female to male ratio were 17.5:1.Coexisting epilepsy was present in 6 (8.10%).40 patients(54.05%) had received one or multiple antiepileptic drugs. Out of 74 patients of PNES 48(64.86%) had predominant motor phenomenon, whereas 26 pa...
ArXiv, 2021
Knowledge Base Question Answering (KBQA) tasks that involve complex reasoning are emerging as an ... more Knowledge Base Question Answering (KBQA) tasks that involve complex reasoning are emerging as an important research direction. However, most KBQA systems struggle with generalizability, particularly on two dimensions: (a) across multiple reasoning types where both datasets and systems have primarily focused on multi-hop reasoning, and (b) across multiple knowledge bases, where KBQA approaches are specifically tuned to a single knowledge base. In this paper, we present SYGMA, a modular approach facilitating generalizability across multiple knowledge bases and multiple reasoning types. Specifically, SYGMA contains three high level modules: 1) KB-agnostic question understanding module that is common across KBs 2) Rules to support additional reasoning types and 3) KB-specific question mapping and answering module to address the KB-specific aspects of the answer extraction. We demonstrate effectiveness of our system by evaluating on datasets belonging to two distinct knowledge bases, DBp...
Knowledge Graph Question Answering (KGQA) has become a prominent area in natural language process... more Knowledge Graph Question Answering (KGQA) has become a prominent area in natural language processing due to the emergence of large-scale Knowledge Graphs (KGs). Recently Neural Machine Translation based approaches are gaining momentum that translates natural language queries to structured query languages thereby solving the KGQA task. However, most of these methods struggle with out-of-vocabulary words where test entities and relations are not seen during training time. In this work, we propose a modular two-stage neural architecture to solve the KGQA task. The first stage generates a sketch of the target SPARQL called SPARQL silhouette for the input question. This comprises of (1) Noise simulator to facilitate out-of-vocabulary words and to reduce vocabulary size (2) seq2seq model for text to SPARQL silhouette generation. The second stage is a Neural Graph Search Module. SPARQL silhouette generated in the first stage is distilled in the second stage by substituting precise relation...
Knowledge base question answering (KBQA) is an important task in Natural Language Processing. Exi... more Knowledge base question answering (KBQA) is an important task in Natural Language Processing. Existing approaches face significant challenges including complex question understanding, necessity for reasoning, and lack of large end-to-end training datasets. In this work, we propose Neuro-Symbolic Question Answering (NSQA), a modular KBQA system, that leverages (1) Abstract Meaning Representation (AMR) parses for task-independent question understanding; (2) a simple yet effective graph transformation approach to convert AMR parses into candidate logical queries that are aligned to the KB; (3) a pipeline-based approach which integrates multiple, reusable modules that are trained specifically for their individual tasks (semantic parser, entity and relationship linkers, and neuro-symbolic reasoner) and do not require end-to-end training data. NSQA achieves state-of-the-art performance on two prominent KBQA datasets based on DBpedia (QALD-9 and LC-QuAD 1.0). Furthermore, our analysis emph...
Noncompressive myelopathy of lower dorsal spine secondary to trauma is a rare event. We report a ... more Noncompressive myelopathy of lower dorsal spine secondary to trauma is a rare event. We report a case of delayed paraplegia in a patient with a history of road traffic accident. The X-ray of dorsolumbar spine did not show any abnormality. Magnetic resonance imaging of dorsolumbar spine was performed which showed the presence of central T2-weighted hyperintensities from D10–D11 to D12–L1 level. No associated bony injury was documented, and the integrity of the spinal canal was maintained. The patient was managed conservatively with bed rest, and steroids were given. However, the patient did not show any signs of improvement after 1 month of follow-up.
Journal of the Neurological Sciences
Neurology and Clinical Neuroscience
To assess retinal nerve fiber layer thickness by optical coherence tomography (OCT) in various ce... more To assess retinal nerve fiber layer thickness by optical coherence tomography (OCT) in various central nervous system demyelinating diseases.
Annals of Movement Disorders
Syringomyelia is described as a fluid-filled cavity within the spinal cord devoid of an ependymal... more Syringomyelia is described as a fluid-filled cavity within the spinal cord devoid of an ependymal lining. It is best visualized on magnetic resonance imaging (MRI), with the cavity being low intensity on T1-weighted and high intensity on T2-weighted images. The association of syringomyelia with dystonia has been infrequently reported in the medical literature. We herein describe a case of syringomyelia with Chiari 1 malformation and hydrocephalus having writer’s cramp and pseudoathetosis, which as per our review, is a yet undescribed manifestation in previously published literature. Also, we emphasize on the usefulness of spinal MRI in a case of focal dystonia/writer’s cramp if cause is not apparent after initial evaluation and more so if associated with proprioceptive sensory impairment and pseudoathetosis.
Proceedings of the AAAI Conference on Artificial Intelligence
Research on the task of Reading Comprehension style Question Answering (RCQA) has gained momentum... more Research on the task of Reading Comprehension style Question Answering (RCQA) has gained momentum in recent years due to the emergence of human annotated datasets and associated leaderboards, for example CoQA, HotpotQA, SQuAD, TriviaQA, etc. While state-of-the-art has advanced considerably, there is still ample opportunity to advance it further on some important variants of the RCQA task. In this paper, we propose a novel deep neural architecture, called TAP (Translucent Answer Prediction), to identify answers and evidence (in the form of supporting facts) in an RCQA task requiring multi-hop reasoning. TAP comprises two loosely coupled networks – Local and Global Interaction eXtractor (LoGIX) and Answer Predictor (AP). LoGIX predicts supporting facts, whereas AP consumes these predicted supporting facts to predict the answer span. The novel design of LoGIX is inspired by two key design desiderata – local context and global interaction– that we identified by analyzing examples of mul...
Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics
We introduce TECHQA, a domain-adaptation question answering dataset for the technical support dom... more We introduce TECHQA, a domain-adaptation question answering dataset for the technical support domain. The TECHQA corpus highlights two real-world issues from the automated customer support domain. First, it contains actual questions posed by users on a technical forum, rather than questions generated specifically for a competition or a task. Second, it has a real-world size-600 training, 310 dev, and 490 evaluation question/answer pairs-thus reflecting the cost of creating large labeled datasets with actual data. Consequently, TECHQA is meant to stimulate research in domain adaptation rather than being a resource to build QA systems from scratch. The dataset was obtained by crawling the IBM Developer and IBM DeveloperWorks forums for questions with accepted answers that appear in a published IBM Technote-a technical document that addresses a specific technical issue. We also release a collection of the 801,998 publicly available Technotes as of April 4, 2019 as a companion resource that might be used for pretraining, to learn representations of the IT domain language.
Pattern Recognition Letters
Many machine learning applications such as in vision, biology and social networking deal with dat... more Many machine learning applications such as in vision, biology and social networking deal with data in high dimensions. Feature selection is typically employed to select a subset of features which improves generalization accuracy as well as reduces the computational cost of learning the model. One of the criteria used for feature selection is to jointly minimize the redundancy and maximize the relevance of the selected features. In this paper, we formulate the task of feature selection as a one class SVM problem in a space where features correspond to the data points and instances correspond to the dimensions. The goal is to look for a representative subset of the features (support vectors) which describes the boundary for the region where the set of the features (data points) exists. This leads to a joint optimization of relevance and redundancy in a principled max-margin framework. Additionally, our formulation enables us to leverage existing techniques for optimizing the SVM objective resulting in highly computationally efficient solutions for the task of feature selection. Specifically, we employ the dual coordinate descent algorithm (Hsieh et al., 2008), originally proposed for SVMs, for our formulation. We use a sparse representation to deal with data in very high dimensions. Experiments on seven publicly available benchmark datasets from a variety of domains show that our approach results in orders of magnitude faster solutions even while retaining the same level of accuracy compared to the state of the art feature selection techniques.
Uploads
Papers by Dinesh Khandelwal