Lakshay Sharma - Academia.edu (original) (raw)
Papers by Lakshay Sharma
2022 IEEE International Conference on Big Data (Big Data)
Cornell University - arXiv, Oct 12, 2022
Cornell University - arXiv, Sep 30, 2022
A key challenge in off-road navigation is that even visually similar or semantically identical te... more A key challenge in off-road navigation is that even visually similar or semantically identical terrain may have substantially different traction properties. Existing work typically assumes a nominal or expected robot dynamical model for planning, which can lead to degraded performance if the assumed models are not realizable given the terrain properties. In contrast, this work introduces a new probabilistic representation of traversability as a distribution of parameters in the robot's dynamical model that are conditioned on the terrain characteristics. This model is learned in a selfsupervised manner by fitting a probability distribution over the parameters identified online, encoded as a neural network that takes terrain features as input. This work then presents two risk-aware planning algorithms that leverage the learned traversability model to plan risk-aware trajectories. Finally, a method for detecting unfamiliar terrain with respect to the training data is introduced based on a Gaussian Mixture Model fit to the latent space of the trained model. Experiments demonstrate that the proposed approach outperforms existing work that assumes nominal or expected robot dynamics in both success rate and completion time for representative navigation tasks. Furthermore, when the proposed approach is deployed in an unseen environment, excluding unfamiliar terrains during planning leads to improved success rate.
Cornell University - arXiv, Dec 3, 2017
In this project we analysed how much semantic information images carry, and how much value image ... more In this project we analysed how much semantic information images carry, and how much value image data can add to sentiment analysis of the text associated with the images. To better understand the contribution from images, we compared models which only made use of image data, models which only made use of text data, and models which combined both data types. We also analysed if this approach could help sentiment classifiers generalize to unknown sentiments.
Annals of the Romanian Society for Cell Biology, Mar 1, 2021
The Astrophysical Journal , 247 (1) , Article 32. (2020), Mar 1, 2020
We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("... more We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("Y4" data) of the Dark Energy Survey (DES). The survey covers a contiguous 5000 deg 2 of the southern sky in the grizY optical/NIR filter set, with a typical TNO in this part of the sky being targeted by 25-30 Y4 exposures. This paper focuses on the methods used to detect these objects from the ≈60,000 Y4 exposures, a process made challenging by the absence of the few-hour repeat observations employed by TNO-optimized surveys. Newly developed techniques include: transient/moving object detection by comparison of single-epoch catalogs to catalogs of "stacked" images; quantified astrometric error from atmospheric turbulence; new software for detecting TNO linkages in a temporally sparse transient catalog, and for estimating the rate of spurious linkages; use of faint stars to determine the detection efficiency versus magnitude in all exposures. Final validation of the reality of linked orbits uses a new "sub-threshold confirmation" test, wherein we demand the object be detectable in a stack of the exposures in which the orbit indicates an object should be present, but was not individually detected. This catalog contains all validated TNOs which were detected on 6 unique nights in the Y4 data, and is complete to r23.3 mag with virtually no dependence on orbital properties for bound TNOs at distance 30 au<d<2500 au. The catalog includes 245 discoveries by DES, 139 not previously published. The final DES TNO catalog is expected to yield >0.3 mag more depth, and arcs of >4 yr for nearly all detections.
2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2019
Spatial cognition, as gained through the sense of vision, is one of the most important capabiliti... more Spatial cognition, as gained through the sense of vision, is one of the most important capabilities of human beings. However, for the visually impaired (VI), lack of this perceptual capability poses great challenges in their life. Therefore, we have designed Point-to-Tell-and-Touch, a wearable system with an ergonomic human-machine interface, for assisting the VI with active environmental exploration, with a particular focus on spatial intelligence and navigation to objects of interest in an alien environment. Our key idea is to link visual signals, as decoded synthetically, to the VI’s proprioception for more intelligible guidance, in addition to vision-to-audio assistance, i.e., finger pose, as indicated by pointing, is used as “proprioceptive laser pointer” to target an object in that line of sight. The whole system consists of two features, Point-to-Tell and Point-to-Touch, both of which can work independently or cooperatively. The Point-to-Tell feature contains a camera with a novel one-stage neural network tailored for blind-centered object detection and recognition, and a headphone telling the VI the semantic label and distance from the pointed object. the Point-to-Touch, the second feature, leverages a vibrating wrist band to create a haptic feedback tool that supplements the initial vectorial guidance provided by the first stage (hand pose being direction and the distance being the extent, offered through audio cues). Both platform features utilize proprioception or joint position sense. Through hand pose, the VI end user knows where he or she is pointing relative to their egocentric coordinate system and we are able to use this foundation to build spatial intelligence. Our successful indoor experiments demonstrate the proposed system to be effective and reliable in helping the VI gain spatial cognition and explore the world in a more intuitive way.
SSRN Electronic Journal, 2022
There has been several technical researches and engineering programs being carried out generally ... more There has been several technical researches and engineering programs being carried out generally in an effort to monitor environmental deprivation, natural and man-made calamities. These are part of a process to forecast possible terrible events and help minimize their shocking impact on the environment and human life.
ArXiv, 2019
This paper explores the task Natural Language Understanding (NLU) by looking at duplicate questio... more This paper explores the task Natural Language Understanding (NLU) by looking at duplicate question detection in the Quora dataset. We conducted extensive exploration of the dataset and used various machine learning models, including linear and tree-based models. Our final finding was that a simple Continuous Bag of Words neural network model had the best performance, outdoing more complicated recurrent and attention based models. We also conducted error analysis and found some subjectivity in the labeling of the dataset.
ArXiv, 2019
In recent years, the biggest advances in major Computer Vision tasks, such as object recognition,... more In recent years, the biggest advances in major Computer Vision tasks, such as object recognition, handwritten-digit identification, facial recognition, and many others., have all come through the use of Convolutional Neural Networks (CNNs). Similarly, in the domain of Natural Language Processing, Recurrent Neural Networks (RNNs), and Long Short Term Memory networks (LSTMs) in particular, have been crucial to some of the biggest breakthroughs in performance for tasks such as machine translation, part-of-speech tagging, sentiment analysis, and many others. These individual advances have greatly benefited tasks even at the intersection of NLP and Computer Vision, and inspired by this success, we studied some existing neural image captioning models that have proven to work well. In this work, we study some existing captioning models that provide near state-of-the-art performances, and try to enhance one such model. We also present a simple image captioning model that makes use of a CNN,...
The Astrophysical Journal Supplement Series, 2020
We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("... more We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("Y4" data) of the Dark Energy Survey (DES). The survey covers a contiguous 5000 deg 2 of the southern sky in the grizY optical/NIR filter set, with a typical TNO in this part of the sky being targeted by 25-30 Y4 exposures. This paper focuses on the methods used to detect these objects from the ≈60,000 Y4 exposures, a process made challenging by the absence of the few-hour repeat observations employed by TNO-optimized surveys. Newly developed techniques include: transient/moving object detection by comparison of single-epoch catalogs to catalogs of "stacked" images; quantified astrometric error from atmospheric turbulence; new software for detecting TNO linkages in a temporally sparse transient catalog, and for estimating the rate of spurious linkages; use of faint stars to determine the detection efficiency versus magnitude in all exposures. Final validation of the reality of linked orbits uses a new "sub-threshold confirmation" test, wherein we demand the object be detectable in a stack of the exposures in which the orbit indicates an object should be present, but was not individually detected. This catalog contains all validated TNOs which were detected on 6 unique nights in the Y4 data, and is complete to r23.3 mag with virtually no dependence on orbital properties for bound TNOs at distance 30 au<d<2500 au. The catalog includes 245 discoveries by DES, 139 not previously published. The final DES TNO catalog is expected to yield >0.3 mag more depth, and arcs of >4 yr for nearly all detections.
International Research Journal Of Pharmacy, 2018
Many plants are now been exploited for their medical and pharmacological properties. Acacia sinua... more Many plants are now been exploited for their medical and pharmacological properties. Acacia sinuata has been reported with cure for infectious disease and organ specific disease, confirmed to have analgesic activity and also used to cure jaundice and malarial fever in Ayurveda. Similarly, Adenanthera pavonina is reported for treating wounds, boils, rheumatism, arthritis, diarrhea, and even leprosy. The plant has emetic properties and can induce a person to vomit, but literature studies showed that the safety assessment of such widely used medicinal plants has not been done hence they were chosen for the purpose. The plant extracts of Acacia sinuata and Adenanthera pavonina were tested using chromosomal aberration assay and comet assay for their genotoxicity and further studies were performed to check for the genome repair capability of the plants. The lymphocyte culture was exposed to cisplatin to induce damage to the genome and then treated with the plant extract. In this paper the extracts of A. sinuata and A. pavonina at different concentration significantly increased the mitotic index compared to the cisplatin treated cell alone; this decrease in cell proliferation may be explained by permitting the repair of cisplatin induced DNA damage. Both plants show little genomic activity with increased concentration. All tested concentrations of extract from A. sinuata and A. pavonina had no genotoxic effect on the human blood in vitro.
International Journal of Advanced Research, 2017
Abstract-OVSF codes in the WCDMA system provide facility for variable data rate support. The appl... more Abstract-OVSF codes in the WCDMA system provide facility for variable data rate support. The applications with different rates must be served in fair manner. The paper propose a fair single code and multi code design so that the OVSF code tree is not over-served for single type ...
2022 IEEE International Conference on Big Data (Big Data)
Cornell University - arXiv, Oct 12, 2022
Cornell University - arXiv, Sep 30, 2022
A key challenge in off-road navigation is that even visually similar or semantically identical te... more A key challenge in off-road navigation is that even visually similar or semantically identical terrain may have substantially different traction properties. Existing work typically assumes a nominal or expected robot dynamical model for planning, which can lead to degraded performance if the assumed models are not realizable given the terrain properties. In contrast, this work introduces a new probabilistic representation of traversability as a distribution of parameters in the robot's dynamical model that are conditioned on the terrain characteristics. This model is learned in a selfsupervised manner by fitting a probability distribution over the parameters identified online, encoded as a neural network that takes terrain features as input. This work then presents two risk-aware planning algorithms that leverage the learned traversability model to plan risk-aware trajectories. Finally, a method for detecting unfamiliar terrain with respect to the training data is introduced based on a Gaussian Mixture Model fit to the latent space of the trained model. Experiments demonstrate that the proposed approach outperforms existing work that assumes nominal or expected robot dynamics in both success rate and completion time for representative navigation tasks. Furthermore, when the proposed approach is deployed in an unseen environment, excluding unfamiliar terrains during planning leads to improved success rate.
Cornell University - arXiv, Dec 3, 2017
In this project we analysed how much semantic information images carry, and how much value image ... more In this project we analysed how much semantic information images carry, and how much value image data can add to sentiment analysis of the text associated with the images. To better understand the contribution from images, we compared models which only made use of image data, models which only made use of text data, and models which combined both data types. We also analysed if this approach could help sentiment classifiers generalize to unknown sentiments.
Annals of the Romanian Society for Cell Biology, Mar 1, 2021
The Astrophysical Journal , 247 (1) , Article 32. (2020), Mar 1, 2020
We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("... more We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("Y4" data) of the Dark Energy Survey (DES). The survey covers a contiguous 5000 deg 2 of the southern sky in the grizY optical/NIR filter set, with a typical TNO in this part of the sky being targeted by 25-30 Y4 exposures. This paper focuses on the methods used to detect these objects from the ≈60,000 Y4 exposures, a process made challenging by the absence of the few-hour repeat observations employed by TNO-optimized surveys. Newly developed techniques include: transient/moving object detection by comparison of single-epoch catalogs to catalogs of "stacked" images; quantified astrometric error from atmospheric turbulence; new software for detecting TNO linkages in a temporally sparse transient catalog, and for estimating the rate of spurious linkages; use of faint stars to determine the detection efficiency versus magnitude in all exposures. Final validation of the reality of linked orbits uses a new "sub-threshold confirmation" test, wherein we demand the object be detectable in a stack of the exposures in which the orbit indicates an object should be present, but was not individually detected. This catalog contains all validated TNOs which were detected on 6 unique nights in the Y4 data, and is complete to r23.3 mag with virtually no dependence on orbital properties for bound TNOs at distance 30 au<d<2500 au. The catalog includes 245 discoveries by DES, 139 not previously published. The final DES TNO catalog is expected to yield >0.3 mag more depth, and arcs of >4 yr for nearly all detections.
2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2019
Spatial cognition, as gained through the sense of vision, is one of the most important capabiliti... more Spatial cognition, as gained through the sense of vision, is one of the most important capabilities of human beings. However, for the visually impaired (VI), lack of this perceptual capability poses great challenges in their life. Therefore, we have designed Point-to-Tell-and-Touch, a wearable system with an ergonomic human-machine interface, for assisting the VI with active environmental exploration, with a particular focus on spatial intelligence and navigation to objects of interest in an alien environment. Our key idea is to link visual signals, as decoded synthetically, to the VI’s proprioception for more intelligible guidance, in addition to vision-to-audio assistance, i.e., finger pose, as indicated by pointing, is used as “proprioceptive laser pointer” to target an object in that line of sight. The whole system consists of two features, Point-to-Tell and Point-to-Touch, both of which can work independently or cooperatively. The Point-to-Tell feature contains a camera with a novel one-stage neural network tailored for blind-centered object detection and recognition, and a headphone telling the VI the semantic label and distance from the pointed object. the Point-to-Touch, the second feature, leverages a vibrating wrist band to create a haptic feedback tool that supplements the initial vectorial guidance provided by the first stage (hand pose being direction and the distance being the extent, offered through audio cues). Both platform features utilize proprioception or joint position sense. Through hand pose, the VI end user knows where he or she is pointing relative to their egocentric coordinate system and we are able to use this foundation to build spatial intelligence. Our successful indoor experiments demonstrate the proposed system to be effective and reliable in helping the VI gain spatial cognition and explore the world in a more intuitive way.
SSRN Electronic Journal, 2022
There has been several technical researches and engineering programs being carried out generally ... more There has been several technical researches and engineering programs being carried out generally in an effort to monitor environmental deprivation, natural and man-made calamities. These are part of a process to forecast possible terrible events and help minimize their shocking impact on the environment and human life.
ArXiv, 2019
This paper explores the task Natural Language Understanding (NLU) by looking at duplicate questio... more This paper explores the task Natural Language Understanding (NLU) by looking at duplicate question detection in the Quora dataset. We conducted extensive exploration of the dataset and used various machine learning models, including linear and tree-based models. Our final finding was that a simple Continuous Bag of Words neural network model had the best performance, outdoing more complicated recurrent and attention based models. We also conducted error analysis and found some subjectivity in the labeling of the dataset.
ArXiv, 2019
In recent years, the biggest advances in major Computer Vision tasks, such as object recognition,... more In recent years, the biggest advances in major Computer Vision tasks, such as object recognition, handwritten-digit identification, facial recognition, and many others., have all come through the use of Convolutional Neural Networks (CNNs). Similarly, in the domain of Natural Language Processing, Recurrent Neural Networks (RNNs), and Long Short Term Memory networks (LSTMs) in particular, have been crucial to some of the biggest breakthroughs in performance for tasks such as machine translation, part-of-speech tagging, sentiment analysis, and many others. These individual advances have greatly benefited tasks even at the intersection of NLP and Computer Vision, and inspired by this success, we studied some existing neural image captioning models that have proven to work well. In this work, we study some existing captioning models that provide near state-of-the-art performances, and try to enhance one such model. We also present a simple image captioning model that makes use of a CNN,...
The Astrophysical Journal Supplement Series, 2020
We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("... more We present a catalog of 316 trans-Neptunian bodies (TNOs) detected from the first four seasons ("Y4" data) of the Dark Energy Survey (DES). The survey covers a contiguous 5000 deg 2 of the southern sky in the grizY optical/NIR filter set, with a typical TNO in this part of the sky being targeted by 25-30 Y4 exposures. This paper focuses on the methods used to detect these objects from the ≈60,000 Y4 exposures, a process made challenging by the absence of the few-hour repeat observations employed by TNO-optimized surveys. Newly developed techniques include: transient/moving object detection by comparison of single-epoch catalogs to catalogs of "stacked" images; quantified astrometric error from atmospheric turbulence; new software for detecting TNO linkages in a temporally sparse transient catalog, and for estimating the rate of spurious linkages; use of faint stars to determine the detection efficiency versus magnitude in all exposures. Final validation of the reality of linked orbits uses a new "sub-threshold confirmation" test, wherein we demand the object be detectable in a stack of the exposures in which the orbit indicates an object should be present, but was not individually detected. This catalog contains all validated TNOs which were detected on 6 unique nights in the Y4 data, and is complete to r23.3 mag with virtually no dependence on orbital properties for bound TNOs at distance 30 au<d<2500 au. The catalog includes 245 discoveries by DES, 139 not previously published. The final DES TNO catalog is expected to yield >0.3 mag more depth, and arcs of >4 yr for nearly all detections.
International Research Journal Of Pharmacy, 2018
Many plants are now been exploited for their medical and pharmacological properties. Acacia sinua... more Many plants are now been exploited for their medical and pharmacological properties. Acacia sinuata has been reported with cure for infectious disease and organ specific disease, confirmed to have analgesic activity and also used to cure jaundice and malarial fever in Ayurveda. Similarly, Adenanthera pavonina is reported for treating wounds, boils, rheumatism, arthritis, diarrhea, and even leprosy. The plant has emetic properties and can induce a person to vomit, but literature studies showed that the safety assessment of such widely used medicinal plants has not been done hence they were chosen for the purpose. The plant extracts of Acacia sinuata and Adenanthera pavonina were tested using chromosomal aberration assay and comet assay for their genotoxicity and further studies were performed to check for the genome repair capability of the plants. The lymphocyte culture was exposed to cisplatin to induce damage to the genome and then treated with the plant extract. In this paper the extracts of A. sinuata and A. pavonina at different concentration significantly increased the mitotic index compared to the cisplatin treated cell alone; this decrease in cell proliferation may be explained by permitting the repair of cisplatin induced DNA damage. Both plants show little genomic activity with increased concentration. All tested concentrations of extract from A. sinuata and A. pavonina had no genotoxic effect on the human blood in vitro.
International Journal of Advanced Research, 2017
Abstract-OVSF codes in the WCDMA system provide facility for variable data rate support. The appl... more Abstract-OVSF codes in the WCDMA system provide facility for variable data rate support. The applications with different rates must be served in fair manner. The paper propose a fair single code and multi code design so that the OVSF code tree is not over-served for single type ...