Author name cluster

Arnab Ghosh

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

3 papers

1 author row

NeurIPS Conference 2020 Conference Paper

STEER : Simple Temporal Regularization For Neural ODE

Arnab Ghosh
Harkirat Behl
Emilien Dupont
Philip Torr
Vinay Namboodiri

Training Neural Ordinary Differential Equations (ODEs) is often computationally expensive. Indeed, computing the forward pass of such models involves solving an ODE which can become arbitrarily complex during training. Recent works have shown that regularizing the dynamics of the ODE can partially alleviate this. In this paper we propose a new regularization technique: randomly sampling the end time of the ODE during training. The proposed regularization is simple to implement, has negligible overhead and is effective across a wide variety of tasks. Further, the technique is orthogonal to several other methods proposed to regularize the dynamics of ODEs and as such can be used in conjunction with them. We show through experiments on normalizing flows, time series models and image recognition that the proposed regularization can significantly decrease training time and even improve performance over baseline models.

PDF Details

AAAI Conference 2017 Conference Paper

Contextual RNN-GANs for Abstract Reasoning Diagram Generation

Viveka Kulharia
Arnab Ghosh
Amitabha Mukerjee
Vinay Namboodiri
Mohit Bansal

Understanding, predicting, and generating object motions and transformations is a core problem in artiﬁcial intelligence. Modeling sequences of evolving images may provide better representations and models of motion and may ultimately be used for forecasting, simulation, or video generation. Diagrammatic Abstract Reasoning is an avenue in which diagrams evolve in complex patterns and one needs to infer the underlying pattern sequence and generate the next image in the sequence. For this, we develop a novel Contextual Generative Adversarial Network based on Recurrent Neural Networks (Context-RNN-GANs), where both the generator and the discriminator modules are based on contextual history (modeled as RNNs) and the adversarial discriminator guides the generator to produce realistic images for the particular time step in the image sequence. We evaluate the Context-RNN-GAN model (and its variants) on a novel dataset of Diagrammatic Abstract Reasoning, where it performs competitively with 10th-grade human performance but there is still scope for interesting improvements as compared to college-grade human performance. We also evaluate our model on a standard video next-frame prediction task, achieving improved performance over comparable state-of-the-art.

PDF Details

YNIMG Journal 2014 Journal Article

Cytochrome c oxidase response to changes in cerebral oxygen delivery in the adult brain shows higher brain-specificity than haemoglobin

Christina Kolyva
Arnab Ghosh
Ilias Tachtsidis
David Highton
Chris E. Cooper
Martin Smith
Clare E. Elwell

The redox state of cerebral mitochondrial cytochrome c oxidase monitored with near-infrared spectroscopy (Δ[oxCCO]) is a signal with strong potential as a non-invasive, bedside biomarker of cerebral metabolic status. We hypothesised that the higher mitochondrial density of brain compared to skin and skull would lead to evidence of brain-specificity of the Δ[oxCCO] signal when measured with a multi-distance near-infrared spectroscopy (NIRS) system. Measurements of Δ[oxCCO] as well as of concentration changes in oxygenated (Δ[HbO2]) and deoxygenated haemoglobin (Δ[HHb]) were taken at multiple source-detector distances during systemic hypoxia and hypocapnia (decrease in cerebral oxygen delivery), and hyperoxia and hypercapnia (increase in cerebral oxygen delivery) from 15 adult healthy volunteers. Increasing source-detector spacing is associated with increasing light penetration depth and thus higher sensitivity to cerebral changes. An increase in Δ[oxCCO] was observed during the challenges that increased cerebral oxygen delivery and the opposite was observed when cerebral oxygen delivery decreased. A consistent pattern of statistically significant increasing amplitude of the Δ[oxCCO] response with increasing light penetration depth was observed in all four challenges, a behaviour that was distinctly different from that of the haemoglobin chromophores, which did not show this statistically significant depth gradient. This depth-dependence of the Δ[oxCCO] signal corroborates the notion of higher concentrations of CCO being present in cerebral tissue compared to extracranial components and highlights the value of NIRS-derived Δ[oxCCO] as a brain-specific signal of cerebral metabolism, superior in this aspect to haemoglobin.

Details DOI