Search

Explainable AI (XAI) in medical histopathology is essential for enhancing the interpretability and clinical trustworthiness of deep learning models in cancer diagnosis. However, the black-box nature of these models often limits their clinical adoption. We introduce GRAPHITE (Graph-based Interpretable Tissue Examination), a post-hoc explainable framework designed for breast cancer tissue microarray (TMA) analysis. GRAPHITE employs a multiscale approach, extracting patches at various magnification levels, constructing an hierarchical graph, and utilising graph attention networks (GAT) with scalewise attention (SAN) to capture scale-dependent features. We trained the model on 140 tumour TMA cores and four benign whole slide images from which 140 benign samples were created, and tested it on 53 pathologist-annotated TMA samples. GRAPHITE outperformed traditional XAI methods, achieving a mean average precision (mAP) of 0.56, an area under the receiver operating characteristic curve (AUROC) of 0.94, and a threshold robustness (ThR) of 0.70, indicating that the model maintains high performance across a wide range of thresholds. In clinical utility, GRAPHITE achieved the highest area under the decision curve (AUDC) of 4.17e+5, indicating reliable decision support across thresholds. These results highlight GRAPHITE's potential as a clinically valuable tool in computational pathology, providing interpretable visualisations that align with the pathologists' diagnostic reasoning and support precision medicine., Comment: 24 Pages, 9 Figures, 1 Tables

It has been suggested that black hole microstates in string theory may have "no interiors". The arguments of arXiv:2312.14108 suggest that even if this were the case, conventional bulk EFT should be valid at the horizon for black holes that are not too old. Here, we explore the validity of bulk EFT for young black holes from a different perspective, by considering the gravitational collapse of scalar moduli. This builds on arXiv:2003.05488, which studied the numerical collapse of a scalar field in a Choptuik-like set up. Evidence was presented there for two observations -- (a) For every scalar profile that results in collapse, the scalar field undergoes trans-Planckian variation, (b) Every trans-Planckian scalar motion is hidden behind an apparent horizon. The results of arXiv:2003.05488 were mostly obtained in regimes close to critical collapse, where scale-separation was difficult to achieve. In this paper, we study the same system in super-critical regimes. This allows us to study the localization properties of scalar variations inside the apparent horizon by separating the horizon and singularity scales. We find strong evidence that the $\mathcal{O}(1)$ scalar variation is localized around the curvature divergence, and is hierarchically well-separated from the apparent horizon. Because $\mathcal{O}(1)$ scalar movement is associated to the breakdown of bulk EFT according to the swampland distance conjecture, this is an indication that bulk EFT is well-behaved in the interiors of young black holes. Our results suggest the perspective that cosmic censorship is a mechanism for preserving bulk EFT between the horizon and the singularity., Comment: 16 pages, 16 plots

Light thermal dark matter has emerged as an attractive theoretical possibility and a promising target for discovery at experiments in the near future. Such scenarios generically invoke mediators with very small couplings to the Standard Model, but moderately strong couplings within the dark sector, calling into question theoretical estimates based on the lowest order of perturbation theory. As an example, we focus on a scenario in which (pseudo)-Dirac fermion dark matter is connected to the standard model via a dark photon charged under a new $U(1)^{\prime}$ extension of the standard model, and we investigate the impact of the next-to-leading order corrections to annihilation and scattering. We find that radiative corrections can significantly impact model predictions for the relic density and scattering cross-section, depending on the strength of the dark sector coupling and ratio of the dark matter to mediator mass. We also show why factorization into the yield parameter $Y$ typically presented in literature leads to imprecision. Our results are necessary to accurately map experimental searches into the model parameter space and assess their ability to reach thermal production targets., Comment: 10 pages, 5 figures

Survival risk stratification is an important step in clinical decision making for breast cancer management. We propose a novel deep learning approach for this purpose by integrating histopathological imaging, genetic and clinical data. It employs vision transformers, specifically the MaxViT model, for image feature extraction, and self-attention to capture intricate image relationships at the patient level. A dual cross-attention mechanism fuses these features with genetic data, while clinical data is incorporated at the final layer to enhance predictive accuracy. Experiments on the public TCGA-BRCA dataset show that our model, trained using the negative log likelihood loss function, can achieve superior performance with a mean C-index of 0.64, surpassing existing methods. This advancement facilitates tailored treatment strategies, potentially leading to improved patient outcomes., Comment: Keywords: Multimodal Fusion, Breast Cancer, Whole Slide Images, Survival Prediction

Breast cancer is a significant health concern affecting millions of women worldwide. Accurate survival risk stratification plays a crucial role in guiding personalised treatment decisions and improving patient outcomes. Here we present BioFusionNet, a deep learning framework that fuses image-derived features with genetic and clinical data to obtain a holistic profile and achieve survival risk stratification of ER+ breast cancer patients. We employ multiple self-supervised feature extractors (DINO and MoCoV3) pretrained on histopathological patches to capture detailed image features. These features are then fused by a variational autoencoder and fed to a self-attention network generating patient-level features. A co-dual-cross-attention mechanism combines the histopathological features with genetic data, enabling the model to capture the interplay between them. Additionally, clinical data is incorporated using a feed-forward network, further enhancing predictive performance and achieving comprehensive multimodal feature integration. Furthermore, we introduce a weighted Cox loss function, specifically designed to handle imbalanced survival data, which is a common challenge. Our model achieves a mean concordance index of 0.77 and a time-dependent area under the curve of 0.84, outperforming state-of-the-art methods. It predicts risk (high versus low) with prognostic significance for overall survival in univariate analysis (HR=2.99, 95% CI: 1.88--4.78, p<0.005), and maintains independent significance in multivariate analysis incorporating standard clinicopathological variables (HR=2.91, 95\% CI: 1.80--4.68, p<0.005)., Comment: Keywords: Multimodal Fusion, Breast Cancer, Whole Slide Images, Deep Neural Network, Survival Prediction

We continue the study of dipole cosmology framework put forward in \cite{Krishnan:2022qbv}, a beyond FLRW setting that has a preferred direction in the metric which may be associated with a cosmological tilt, a cosmic dipole. In this setup the shear and the tilt can be positive or negative given the dipole direction. We thoroughly analyze evolution of the universe in this setting, particularly focusing on the behaviour near the Big Bang (BB). We first analyze a single fluid model with a generic constant equation of state $w$. While details of the behavior near the BB depends on $w$ and the other initial conditions, we find that when the shear is negative we have a shear dominated BB singularity, whereas for a positive shear we have a much milder singularity, the whimper singularity \cite{Ellis:1974ug}, at which the tilt blows up while curvature invariants remain finite. We then consider dipole $\Lambda$CDM model which besides the shear has two tilt parameters, one for radiation and one for the pressureless matter. For positive (negative) shear we again find whimper (curvature) singularity near the BB. Moreover, when the tilt parameters have opposite signs, the shear can change sign from negative to positive in the course of evolution of the Universe. We show that the relative tilt of the radiation and the matter generically remains sizable at late times., Comment: 35 pages, 15 figures, 2 tables

Dipole cosmology is the maximally Copernican generalization of the FLRW paradigm that can incorporate bulk flows in the cosmic fluid. In this paper, we first discuss how multiple fluid components with independent flows can be realized in this set up. This is the necessary step to promote ``tilted" Bianchi cosmologies to a viable framework for cosmological model building involving fluid mixtures (as in FLRW). We present a dipole $\Lambda$CDM model which has radiation and matter with independent flows, with (or without) a positive cosmological constant. A remarkable feature of models containing radiation (including dipole $\Lambda$CDM) is that the $relative$ flow between radiation and matter can increase at late times, which can contribute to eg., the CMB dipole. This can happen generically in the space of initial conditions. We discuss the significance of this observation for late time cosmic tensions., Comment: 35 pages, many figures; v2 references added

Gene expression can be used to subtype breast cancer with improved prediction of risk of recurrence and treatment responsiveness over that obtained using routine immunohistochemistry (IHC). However, in the clinic, molecular profiling is primarily used for ER+ breast cancer, which is costly, tissue destructive, requires specialized platforms and takes several weeks to obtain a result. Deep learning algorithms can effectively extract morphological patterns in digital histopathology images to predict molecular phenotypes quickly and cost-effectively. We propose a new, computationally efficient approach called hist2RNA inspired by bulk RNA-sequencing techniques to predict the expression of 138 genes (incorporated from six commercially available molecular profiling tests), including luminal PAM50 subtype, from hematoxylin and eosin (H&E) stained whole slide images (WSIs). The training phase involves the aggregation of extracted features for each patient from a pretrained model to predict gene expression at the patient level using annotated H&E images from The Cancer Genome Atlas (TCGA, n=335). We demonstrate successful gene prediction on a held-out test set (n = 160, corr = 0.82 across patients, corr = 0.29 across genes) and perform exploratory analysis on an external tissue microarray (TMA) dataset (n = 498) with known IHC and survival information. Our model is able to predict gene expression and luminal PAM50 subtype (Luminal A versus Luminal B) on the TMA dataset with prognostic significance for overall survival in univariate analysis (c-index = 0.56, hazard ratio = 2.16 (95% CI 1.12-3.06), p < 5 x 10-3), and independent significance in multivariate analysis incorporating standard clinicopathological variables (c-index = 0.65, hazard ratio = 1.85 (95% CI 1.30-2.68), p < 5 x 10-3)., Comment: 16 pages, 10 figures, 2 tables

The neural implementation of operant conditioning with few trials is unclear. We propose a Hippocampus-Inspired Cognitive Architecture (HICA) as a neural mechanism for operant conditioning. HICA explains a learning mechanism in which agents can learn a new behavior policy in a few trials, as mammals do in operant conditioning experiments. HICA is composed of two different types of modules. One is a universal learning module type that represents a cortical column in the neocortex gray matter. The working principle is modeled as Modulated Heterarchical Prediction Memory (mHPM). In mHPM, each module learns to predict a succeeding input vector given the sequence of the input vectors from lower layers and the context vectors from higher layers. The prediction is fed into the lower layers as a context signal (top-down feedback signaling), and into the higher layers as an input signal (bottom-up feedforward signaling). Rewards modulate the learning rate in those modules to memorize meaningful sequences effectively. In mHPM, each module updates in a local and distributed way compared to conventional end-to-end learning with backpropagation of the single objective loss. This local structure enables the heterarchical network of modules. The second type is an innate, special-purpose module representing various organs of the brain's subcortical system. Modules modeling organs such as the amygdala, hippocampus, and reward center are pre-programmed to enable instinctive behaviors. The hippocampus plays the role of the simulator. It is an autoregressive prediction model of the top-most level signal with a loop structure of memory, while cortical columns are lower layers that provide detailed information to the simulation. The simulation becomes the basis for learning with few trials and the deliberate planning required for operant conditioning., Comment: arXiv admin note: text overlap with arXiv:2108.03793

We show that the Friedmann-Lema\^{i}tre-Robertson-Walker (FLRW) framework has an instability towards the growth of fluid flow anisotropies, even if the Universe is accelerating. This flow (tilt) instability in the matter sector is invisible to Cosmic No-Hair Theorem-like arguments, which typically only flag shear anisotropies in the metric. We illustrate our claims in the setting of ``dipole cosmology'', the maximally Copernican generalization of FLRW that can accommodate a flow. Simple models are sufficient to show that the cosmic flow need not track the shear, even in the presence of a positive cosmological constant. We also emphasize that the growth of the tilt hair is fairly generic if the total equation of state $w(t) \rightarrow -1$ at late times (as it does in standard cosmology), irrespective of the precise model of dark energy. The generality of our theoretical result puts various recent observational claims about late time anisotropies and cosmic dipoles in a new light., Comment: 5 pages, 4 figures

We introduce the $dipole$ $cosmological$ $principle$, the idea that the Universe is a maximally Copernican cosmology, compatible with a cosmic flow. It serves as the most symmetric paradigm that generalizes the FLRW ansatz, in light of the increasingly numerous (but still tentative) hints that have emerged in the last two decades for a non-kinematic component in the CMB dipole. Einstein equations in our "dipole cosmology" are still ordinary differential equations -- but instead of the two Friedmann equations, now we have four. The two new functions can be viewed as an anisotropic scale factor that breaks the isotropy group from $SO(3)$ to $U(1)$, and a "tilt" that captures the cosmic flow velocity. The result is an axially isotropic, tilted Bianchi V/VII$_h$ cosmology. We assess the possibility of model building within the dipole cosmology paradigm, and discuss the dynamics of expansion rate, anisotropic shear and tilt, in various examples. A key observation is that the cosmic flow (tilt) can grow even while the anisotropy (shear) dies down. Remarkably, this can happen even in an era of late time acceleration., Comment: 44 pages, Many Figures

The neutrino sector offers one of the most sensitive probes of new physics beyond the Standard Model of Particle Physics. The mechanism of neutrino mass generation is still unknown. The observed suppression of neutrino masses hints at a large scale, conceivably of the order of the scale of a Grand Unified Theory (GUT), a unique feature of neutrinos that is not shared by the charged fermions. The origin of neutrino masses and mixing is part of the outstanding puzzle of fermion masses and mixings, which is not explained in the SM. Flavor model building for both quark and lepton sectors is important in order to gain a better understanding of the origin of the structure of mass hierarchy and flavor mixing, which constitute the dominant fraction of the SM parameters. Recent activities in neutrino flavor model building based on non-Abelian discrete flavor symmetries and modular flavor symmetries have been shown to be a promising direction to explore. The emerging models provide a framework that has a significantly reduced number of undetermined parameters in the flavor sector. Model building based on non-Abelian discrete flavor symmetries and their modular variants enables the particle physics community to interpret the current and anticipated upcoming data from neutrino experiments. Pursuit of flavor model building based on such frameworks can also provide connections to possible UV completions, in particular to string theory. We emphasize the importance of constructing models in which the uncertainties of theoretical predictions are smaller than, or at most compatible with, the error bars of measurements in neutrino experiments., Comment: 24 pages, 7 figures, Contribution to Snowmass 2021

We reverse the logic behind the apparent existence of $H_0$-tension, to design diagnostics for cosmological models. The basic idea is that the non-constancy of $H_0$ inferred from observations at different redshifts is a null hypothesis test for models within the FLRW paradigm -- if $H_0$ runs, the model is wrong. Depending on the kind of observational data, the most suitable form of the diagnostic can vary. As examples, we present two $H_0$ diagnostics that are adapted to two different BAO observables. We use these and the corresponding BAO data to Gaussian reconstruct the running of $H_0$ in flat $\Lambda$CDM with Planck values for the model parameters. For flat $\Lambda$CDM when the radiation contribution can be neglected, with comoving distance data, the diagnostic is a simple hypergeometric function. Possible late time deviations from the FLRW paradigm can also be accommodated, by simply keeping track of the (potentially anisotropic) sky variation of the diagnostic., Comment: 9 pages, 5 plots

We look at an algorithmic information theory based definition of value of a creative artifact and discuss the computational difficulty associated with the creation or determination of value. We look at the computational resources required to create a valuable product and discuss how likely an observer is to verify a claim that an object required significant effort on the creator's part.

The traditional oil supply chain suffers from various shortcomings regarding crude oil extraction, processing, distribution, environmental pollution, and traceability. It offers an only a forward flow of products with almost no security and tracking process. In time, the system will lag behind due to the limitations in quality inspection, fraudulent information, and monopolistic behavior of supply chain entities. Inclusion of counterfeiting products and opaqueness of the system urge renovation in this sector. The recent evolution of Industry 4.0 leads to the alternation in the supply chain introducing the smart supply chain. Technological advancement can now reshape the infrastructure of the supply chain for the future. In this paper, we suggest a conceptual framework utilizing Blockchain and Smart Contract to monitor the overall oil supply chain. Blockchain is a groundbreaking technology to monitor and support the security building of a decentralized type supply chain over a peer-to-peer network. The use of the Internet of Things (IoT), especially sensors, opens broader window to track the global supply chain in real-time. We construct a methodology to support reverse traceability for each participant of the supply chain. The functions and characteristics of Blockchain and Smart Contract are defined. Implementation of Smart Contracts has also been shown with detailed analysis. We further describe the challenges of implementing such a system and validate our framework's adaptability in the real world. The paper concludes with future research scope to mitigate the restrictions of data management and maintenance with advanced working prototypes and agile systems achieving greater traceability and transparency., Comment: Accepted as Open access article in IET Blockchain

Task-specific AI agents are showing remarkable performance across different domains. But modeling generalized AI agents like human intelligence will require more than current datasets or only reward-based environments that don't include experiences that an infant gathers throughout its initial stages. In this paper, we present Simulated Environment for Developmental Robotics (SEDRo). It simulates the environments for a baby agent that a human baby experiences throughout the pre-born fetus stage to post-birth 12 months. SEDRo also includes a mother character to provide social interaction with the agent. To evaluate different developmental milestones of the agent, SEDRo incorporates some experiments from developmental psychology., Comment: 5 pages ( Including 1 page reference), 3 figures, Presented at BabyMind workshop, NIPS 2020

Despite recent advances of AI research in many application-specific domains, we do not know how to build a human-level artificial intelligence (HLAI). We conjecture that learning from others' experience with the language is the essential characteristic that distinguishes human intelligence from the rest. Humans can update the action-value function with the verbal description as if they experience states, actions, and corresponding rewards sequences firsthand. In this paper, we present a classification of intelligence according to how individual agents learn and propose a definition and a test for HLAI. The main idea is that language acquisition without explicit rewards can be a sufficient test for HLAI.

Washing hands properly and frequently is the simplest and most cost-effective interventions to prevent the spread of infectious diseases. People are often ignorant about proper handwashing in different situations and do not know if they wash hands properly. Smartwatches are found to be effective for assessing the quality of handwashing. However, the existing smartwatch based systems are not comprehensive enough in terms of achieving accuracy as well as reminding people to handwash and providing feedback to the user about the quality of handwashing. On-device processing is often required to provide real-time feedback to the user, and so it is important to develop a system that runs efficiently on low-resource devices like smartwatches. However, none of the existing systems for handwashing quality assessment are optimized for on-device processing. We present iWash, a comprehensive system for quality assessment and context-aware reminder for handwashing with real-time feedback using smartwatches. iWash is a hybrid deep neural network based system that is optimized for on-device processing to ensure high accuracy with minimal processing time and battery usage. Additionally, it is a context-aware system that detects when the user is entering home using a Bluetooth beacon and provides reminders to wash hands. iWash also offers touch-free interaction between the user and the smartwatch that minimizes the risk of germ transmission. We collected a real-life dataset and conducted extensive evaluations to demonstrate the performance of iWash. Compared to the existing handwashing quality assessment systems, we achieve around 12% higher accuracy for quality assessment, as well as we reduce the processing time and battery usage by around 37% and 10%, respectively., Comment: Published in Elsevier Smart Health Journal, 2021

Even with impressive advances in application-specific models, we still lack knowledge about how to build a model that can learn in a human-like way and do multiple tasks. To learn in a human-like way, we need to provide a diverse experience that is comparable to humans. In this paper, we introduce our ongoing effort to build a simulated environment for developmental robotics (SEDRo). SEDRo provides diverse human experiences ranging from those of a fetus to a 12th-month-old. A series of simulated tests based on developmental psychology will be used to evaluate the progress of a learning model. We anticipate SEDRo to lower the cost of entry and facilitate research in the developmental robotics community.

As the current trend of artificial intelligence is shifting towards self-supervised learning, conventional norms such as highly curated domain-specific data, application-specific learning models, extrinsic reward based learning policies etc. might not provide with the suitable ground for such developments. In this paper, we introduce SEDRo, a Simulated Environment for Developmental Robotics which allows a learning agent to have similar experiences that a human infant goes through from the fetus stage up to 12 months. A series of simulated tests based on developmental psychology will be used to evaluate the progress of a learning model., Comment: Presented at Workshop on Learning in Artificial Open Worlds held with ICML 2020

Obesity is a risk factor for many health issues, including heart disease, diabetes, osteoarthritis, and certain cancers. One of the primary behavioral causes, dietary intake, has proven particularly challenging to measure and track. Current behavioral science suggests that family eating dynamics (FED) have high potential to impact child and parent dietary intake, and ultimately the risk of obesity. Monitoring FED requires information about when and where eating events are occurring, the presence or absence of family members during eating events, and some person-level states such as stress, mood, and hunger. To date, there exists no system for real-time monitoring of FED. This paper presents MFED, the first of its kind of system for monitoring FED in the wild in real-time. Smart wearables and Bluetooth beacons are used to monitor and detect eating activities and the location of the users at home. A smartphone is used for the Ecological Momentary Assessment (EMA) of a number of behaviors, states, and situations. While the system itself is novel, we also present a novel and efficient algorithm for detecting eating events from wrist-worn accelerometer data. The algorithm improves eating gesture detection F1-score by 19% with less than 20% computation compared to the state-of-the-art methods. To date, the MFED system has been deployed in 20 homes with a total of 74 participants, and responses from 4750 EMA surveys have been collected. This paper describes the system components, reports on the eating detection results from the deployments, proposes two techniques for improving ground truth collection after the system is deployed, and provides an overview of the FED data, generated from the multi-component system, that can be used to model and more comprehensively understand insights into the monitoring of family eating dynamics.

Optical switches are one of the most important elements of integrated photonics. Here, we designed, fabricated and characterized several nanophotonic optical switches (NOSs) in silicon that exhibit ultra-compact footprint, along with excellent extinction ratios and operating bandwidths. Simulations indicate that our best device of dimensions 3umX3um and 0.3um device height can provide about 10dB extinction ratio with a bandwidth of 12nm, centered at the design wavelength of 1550nm. Three-dimensional finite-difference time-domain analysis in conjunction with a modified version of direct binary search algorithm was used to design the nanophotonic device structure. The modulation is a result of optical modulation of the refractive index of a photochromic material and an optimized geometry of the digital metamaterials that comprise the device. The devices are CMOS fabrication compatible and examples of ones designed using multi-state geometric optimization that will lead to a new class of ultra-compact and multi-functional active integrated-silicon devices., Comment: arXiv admin note: substantial text overlap with arXiv:1712.02835