-
FACTIFY3M: A Benchmark for Multimodal Fact Verification with Explainability through 5W Question-Answering
Authors:
Megha Chakraborty,
Khushbu Pahwa,
Anku Rani,
Shreyas Chatterjee,
Dwip Dalal,
Harshit Dave,
Ritvik G,
Preethi Gurumurthy,
Adarsh Mahor,
Samahriti Mukherjee,
Aditya Pakala,
Ishan Paul,
Janvita Reddy,
Arghya Sarkar,
Kinjal Sensharma,
Aman Chadha,
Amit P. Sheth,
Amitava Das
Abstract:
Combating disinformation is one of the burning societal crises -- about 67% of the American population believes that disinformation produces a lot of uncertainty, and 10% of them knowingly propagate disinformation. Evidence shows that disinformation can manipulate democratic processes and public opinion, causing disruption in the share market, panic and anxiety in society, and even death during cr…
▽ More
Combating disinformation is one of the burning societal crises -- about 67% of the American population believes that disinformation produces a lot of uncertainty, and 10% of them knowingly propagate disinformation. Evidence shows that disinformation can manipulate democratic processes and public opinion, causing disruption in the share market, panic and anxiety in society, and even death during crises. Therefore, disinformation should be identified promptly and, if possible, mitigated. With approximately 3.2 billion images and 720,000 hours of video shared online daily on social media platforms, scalable detection of multimodal disinformation requires efficient fact verification. Despite progress in automatic text-based fact verification (e.g., FEVER, LIAR), the research community lacks substantial effort in multimodal fact verification. To address this gap, we introduce FACTIFY 3M, a dataset of 3 million samples that pushes the boundaries of the domain of fact verification via a multimodal fake news dataset, in addition to offering explainability through the concept of 5W question-answering. Salient features of the dataset include: (i) textual claims, (ii) ChatGPT-generated paraphrased claims, (iii) associated images, (iv) stable diffusion-generated additional images (i.e., visual paraphrases), (v) pixel-level image heatmap to foster image-text explainability of the claim, (vi) 5W QA pairs, and (vii) adversarial fake news stories.
△ Less
Submitted 30 October, 2023; v1 submitted 22 May, 2023;
originally announced June 2023.
-
Mechanisms of drag reduction by semi-dilute inertial particles in turbulent channel flow
Authors:
Himanshu Dave,
M. Houssem Kasbaoui
Abstract:
We investigate the mechanisms by which inertial particles dispersed at semi-dilute conditions cause significant drag-reduction in a turbulent channel flow at $\mathrm{Re}_τ= 180$. We consider a series of four-way coupled Euler-Lagrange simulations where particles having friction Stokes number $\mathrm{St}^+ = 6$ or 30 are introduced at progressively increasing mass loading from $M=0.2$ to 1.0. The…
▽ More
We investigate the mechanisms by which inertial particles dispersed at semi-dilute conditions cause significant drag-reduction in a turbulent channel flow at $\mathrm{Re}_τ= 180$. We consider a series of four-way coupled Euler-Lagrange simulations where particles having friction Stokes number $\mathrm{St}^+ = 6$ or 30 are introduced at progressively increasing mass loading from $M=0.2$ to 1.0. The simulations show that $\mathrm{St}^+ = 30$ particles cause large drag-reduction by up to 19.74\% at $M=1.0$, whereas $\mathrm{St}^+ = 6$ particles cause large drag increase by up to 16.92\% at $M=1.0$. To reveal the mechanisms underpinning drag-reduction or drag-increase, we investigate the stress distribution within the channel and the impact of the dispersed particles on the near-wall coherent structures. We find a distinctive feature of drag-reducing particles which consists in the formation of extremely long clusters, called ropes. These structures align preferentially with the low-speed streaks and contribute to their stabilization and suppression of bursting. Despite the additional stresses due to the particles, the modulation of the near-wall coherent structures leads to a greater reduction of Reynolds shear stresses and partial relaminarization of the near-wall flow. In the case of the drag-increasing particles with $\mathrm{St}^+ = 6$, a reduction in Reynolds shear stresses is also observed, however, this reduction is insufficient to overcome the additional particle stresses which leads to drag increase.
△ Less
Submitted 28 March, 2023;
originally announced March 2023.
-
The volume-filtering immersed boundary method
Authors:
Himanshu Dave,
Marcus Herrmann,
M. Houssem Kasbaoui
Abstract:
We present a novel framework to deal with static and moving immersed boundaries (IB). In this strategy, called Volume-Filtering Immersed Boundary (VFIB) method, transport equations are derived by filtering the Navier-Stokes equations and accounting for stresses at the solid-fluid interface. The result is that boundary conditions that normally apply on the interface are transformed into bodyforces…
▽ More
We present a novel framework to deal with static and moving immersed boundaries (IB). In this strategy, called Volume-Filtering Immersed Boundary (VFIB) method, transport equations are derived by filtering the Navier-Stokes equations and accounting for stresses at the solid-fluid interface. The result is that boundary conditions that normally apply on the interface are transformed into bodyforces that apply on the right-hand side of the filtered transport equations. The filter width acts as a parameter that controls the level of resolution. The IB is considered well-resolved if the filter width is much smaller than the characteristic corrugation scale of the interface. There are several innovations in this IB method. First, it sheds light on the role of the internal flow which arises when the transport equations are solved inside the IB. We show that, it is essential to separate stresses due to the external and internal fluids in order to get accurate forces and provide a method to do so. Second, we show that volumes associated with Lagrangian forcing points on the boundary depend on the local topology of the surface. We provide a method to compute these volumes using a triangle tessellation of the interface and the surface density function. Third, we provide an efficient procedure to compute the solid volume fraction, thus, enabling tagging interior/exterior cells. This volume fraction is also involved in the procedure to separate stresses due to the external fluid from the total stresses. Fourth, we show a path forward to extend the VFIB method to Large Eddy Simulations involving IBs. Lastly, we apply the VFIB in several numerical tests involving two- and three- dimensional static and moving IBs. We show greatly improved results compared to prior IB methods. Further, we test several filter kernels and show that, for well-resolved IBs, the choice of the kernel plays little role.
△ Less
Submitted 25 March, 2023; v1 submitted 30 September, 2022;
originally announced October 2022.
-
Local flame displacement speeds of hydrogen-air premixed flames in moderate to intense turbulence
Authors:
Yuvraj,
Wonsik Song,
Himanshu Dave,
Hong G. Im,
Swetaprovo Chaudhuri
Abstract:
Comprehensive knowledge of local flame displacement speed, $S_d$, in turbulent premixed flames is crucial towards the design and development of hydrogen fuelled next-generation engines. Premixed hydrogen-air flames are characterized by significantly higher laminar flame speed compared to other conventional fuels. Furthermore, in the presence of turbulence, $S_d$ is enhanced much beyond its corresp…
▽ More
Comprehensive knowledge of local flame displacement speed, $S_d$, in turbulent premixed flames is crucial towards the design and development of hydrogen fuelled next-generation engines. Premixed hydrogen-air flames are characterized by significantly higher laminar flame speed compared to other conventional fuels. Furthermore, in the presence of turbulence, $S_d$ is enhanced much beyond its corresponding unstretched, planar laminar value $S_L$. In this study, the effect of high Karlovitz number ($Ka$) turbulence on density-weighted flame displacement speed, $\widetilde{S_d}$, in a H$_2$-air flame is investigated. Recently, it has been identified that flame-flame interactions in regions of large negative curvature govern large deviations of $\widetilde{S_d}$ from $S_L$, for moderately turbulent flames. An interaction model for the same has also been proposed. In this work, we seek to test the interaction model's applicability to intensely turbulent flames characterized by large $Ka$. To that end, we investigate the local flame structures: thermal, chemical structure, the effect of curvature, along the direction that is normal to the chosen isothermal surfaces. Furthermore, relative contributions of the transport and chemistry terms to $\widetilde{S_d}$ are also analyzed. It is found that, unlike the moderately turbulent premixed flames, where enhanced $\widetilde{S_d}$ is driven by interactions among complete flame structures, $\widetilde{S_d}$ enhancement in high $Re_t$ and high $Ka$ flame is predominantly governed by local interactions of the isotherms. It is found that enhancement in $\widetilde{S_d}$ in regions of large negative curvature occurs as a result of these interactions, evincing that the interaction model is useful for high $Ka$ turbulent premixed flames as well.
△ Less
Submitted 15 June, 2021;
originally announced June 2021.
-
Analog Seizure Detection for Implanted Responsive Neurostimulation
Authors:
Abbas A. Zaki,
Noah C. Parker,
Tae-Yoon Kim,
Sam Ishak,
Ty E. Stovall,
Genchang Peng,
Hina Dave,
Jay Harvey,
Mehrdad Nourani,
Xuan Hu,
Alexander J. Edwards,
Joseph S. Friedman
Abstract:
Epilepsy can be treated with medication, however, $30\%$ of epileptic patients are still drug resistive. Devices like responsive neurostimluation systems are implanted in select patients who may not be amenable to surgical resection. However, state-of-the-art devices suffer from low accuracy and high sensitivity. We propose a novel patient-specific seizure detection system based on naïve Bayesian…
▽ More
Epilepsy can be treated with medication, however, $30\%$ of epileptic patients are still drug resistive. Devices like responsive neurostimluation systems are implanted in select patients who may not be amenable to surgical resection. However, state-of-the-art devices suffer from low accuracy and high sensitivity. We propose a novel patient-specific seizure detection system based on naïve Bayesian inference using Müller C-elements. The system improves upon the current leading neurostimulation device, NeuroPace's RNS by implementing analog signal processing for feature extraction, minimizing the power consumption compared to the digital counterpart.
Preliminary simulations were performed in MATLAB, demonstrating that through integrating multiple channels and features, up to $98\%$ detection accuracy for individual patients can be achieved. Similarly, power calculations were performed, demonstrating that the system uses $6.5 μW$ per channel, which when compared to the state-of-the-art NeuroPace system would increase battery life by up to $50 \%$.
△ Less
Submitted 11 June, 2021;
originally announced June 2021.
-
Postoperative brain volumes are associated with one-year neurodevelopmental outcome in children with severe congenital heart disease
Authors:
Eliane Meuwly,
Maria Feldmann,
Walter Knirsch,
Michael von Rhein,
Kelly Payette,
Hitendu Dave,
Ruth Tuura,
Raimund Kottke,
Cornelia Hagmann,
Beatrice Latal,
Andras Jakab
Abstract:
Children with congenital heart disease (CHD) remain at risk for neurodevelopmental impairment despite improved peri- and intraoperative care. Our prospective cohort study aimed to determine the relationship between perioperative brain volumes and neurodevelopmental outcome in neonates with severe CHD. Pre- and postoperative cerebral MRI was acquired in term born neonates with CHD undergoing neonat…
▽ More
Children with congenital heart disease (CHD) remain at risk for neurodevelopmental impairment despite improved peri- and intraoperative care. Our prospective cohort study aimed to determine the relationship between perioperative brain volumes and neurodevelopmental outcome in neonates with severe CHD. Pre- and postoperative cerebral MRI was acquired in term born neonates with CHD undergoing neonatal cardiopulmonary bypass surgery. Brain volumes were measured using an atlas prior based automated method. One-year outcome was assessed with the Bayley-III. CHD infants (n=77) had lower pre- and postoper-ative total and regional brain volumes compared to controls (n=44, all p<0.01). CHD infants had poorer cognitive and motor outcome (p<=0.0001) and a trend towards lower language composite score compared to controls (p=0.06). The total and selected regional postoperative brain volumes predicted cognitive and language outcome (all p<0.04). This association was independent of length of intensive care unit stay for total, cortical, temporal, frontal and cerebellar volumes. In CHD neonates undergoing cardiopulmonary bypass surgery, pre- and postoperative brain volumes are reduced, and postoperative brain volumes predict cognitive and language outcome at one year. Reduced cerebral volumes in CHD patients could serve as a biomarker for im-paired outcome.
△ Less
Submitted 22 January, 2019;
originally announced January 2019.
-
Audio Watermarking with Error Correction
Authors:
Aman Chadha,
Sandeep Gangundi,
Rishabh Goel,
Hiren Dave,
M. Mani Roja
Abstract:
In recent times, communication through the internet has tremendously facilitated the distribution of multimedia data. Although this is indubitably a boon, one of its repercussions is that it has also given impetus to the notorious issue of online music piracy. Unethical attempts can also be made to deliberately alter such copyrighted data and thus, misuse it. Copyright violation by means of unauth…
▽ More
In recent times, communication through the internet has tremendously facilitated the distribution of multimedia data. Although this is indubitably a boon, one of its repercussions is that it has also given impetus to the notorious issue of online music piracy. Unethical attempts can also be made to deliberately alter such copyrighted data and thus, misuse it. Copyright violation by means of unauthorized distribution, as well as unauthorized tampering of copyrighted audio data is an important technological and research issue. Audio watermarking has been proposed as a solution to tackle this issue. The main purpose of audio watermarking is to protect against possible threats to the audio data and in case of copyright violation or unauthorized tampering, authenticity of such data can be disputed by virtue of audio watermarking.
△ Less
Submitted 6 October, 2011;
originally announced October 2011.
