Search

Robust medical Machine Learning (ML) models have the potential to revolutionize healthcare by accelerating clinical research, improving workflows and outcomes, and producing novel insights or capabilities. Developing such ML models from scratch is cost prohibitive and requires substantial compute, data, and time (e.g., expert labeling). To address these challenges, we introduce Health AI Developer Foundations (HAI-DEF), a suite of pre-trained, domain-specific foundation models, tools, and recipes to accelerate building ML for health applications. The models cover various modalities and domains, including radiology (X-rays and computed tomography), histopathology, dermatological imaging, and audio. These models provide domain specific embeddings that facilitate AI development with less labeled data, shorter training times, and reduced computational costs compared to traditional approaches. In addition, we utilize a common interface and style across these models, and prioritize usability to enable developers to integrate HAI-DEF efficiently. We present model evaluations across various tasks and conclude with a discussion of their application and evaluation, covering the importance of ensuring efficacy, fairness, and equity. Finally, while HAI-DEF and specifically the foundation models lower the barrier to entry for ML in healthcare, we emphasize the importance of validation with problem- and population-specific data for each desired usage setting. This technical report will be updated over time as more modalities and features are added., Comment: 16 pages, 8 figures

We investigate the potential for generating an AMPS firewall from the replica wormhole topology, which describes the post-Page time spacetime in the evaporation process of a black hole. Our analysis reveals that this topology gives rise to a Dirac delta force experienced by infalling particles at the event horizon, consistent with the AMPS firewall hypothesis. This force, proportional to the particle's total energy squared, is a direct consequence of the inherent Dirac delta in the Ricci scalar of the replica wormhole. We provide a regularization scheme for this force, which is crucial for obtaining physically meaningful results. Notably, our approach suggests that considering the replica wormhole resolves both the information paradox and the AMPS argument regarding the monogamy of entanglement simultaneously, yielding a more coherent framework., Comment: 7 pages, 2 figures, published in Physics Letters B

Neural decoding, the process of understanding how brain activity corresponds to different stimuli, has been a primary objective in cognitive sciences. Over the past three decades, advancements in functional Magnetic Resonance Imaging and machine learning have greatly improved our ability to map visual stimuli to brain activity, especially in the visual cortex. Concurrently, research has expanded into decoding more complex processes like language and memory across the whole brain, utilizing techniques to handle greater variability and improve signal accuracy. We argue that "seeing" involves more than just mapping visual stimuli onto the visual cortex; it engages the entire brain, as various emotions and cognitive states can emerge from observing different scenes. In this paper, we develop algorithms to enhance our understanding of visual processes by incorporating whole-brain activation maps while individuals are exposed to visual stimuli. We utilize large-scale fMRI encoders and Image generative models pre-trained on large public datasets, which are then fine-tuned through Image-fMRI contrastive learning. Our models hence can decode visual experience across the entire cerebral cortex, surpassing the traditional confines of the visual cortex. We first compare our method with state-of-the-art approaches to decoding visual processing and show improved predictive semantic accuracy by 43%. A network ablation analysis suggests that beyond the visual cortex, the default mode network contributes most to decoding stimuli, in line with the proposed role of this network in sense-making and semantic processing. Additionally, we implemented zero-shot imagination decoding on an extra validation dataset, achieving a p-value of 0.0206 for mapping the reconstructed images and ground-truth text stimuli, which substantiates the model's capability to capture semantic meanings across various scenarios.

Spectroscopic parameters and decays of the exotic tensor meson $T$ with content $bc \overline{b}\overline{c}$ are explored in the context of the diquark-antidiquark model. We treat it as a state built of axial-vector diquark $b^{T}C\gamma _{\mu }c$ and antidiquark $\overline{b}\gamma _{\nu }C \overline{c}^{T}$, where $C$ is the charge conjugation matrix. The mass $m$ and current coupling $\Lambda$ of this tetraquark are extracted from two-point sum rules. Our result for $m=(12.70 \pm 0.09)~\mathrm{GeV}$ proves that $T$ is unstable against strong dissociations to two-meson final states. Its dominant decay channels are processes $T \to J/\psi \Upsilon $, $ \eta_{b}\eta _{c}$, and $B_{c}^{(\ast) +}B_{c}^{(\ast) -}$. Kinematically allowed transformations of $T$ include also decays $T\rightarrow D^{(\ast )+}D^{(\ast )-}$ and $D^{(\ast )0}\overline{D}^{(\ast )0}$, which are generated by $b\overline{b}$ annihilation inside of $T$ and creation later pairs of $DD$ mesons. The full width of $T$ is estimated by considering all of these channels. Their partial widths are calculated by invoking methods of three-point sum rule approach which are required to evaluate strong couplings at corresponding tetraquark-meson-meson vertices. Our predictions for the mass and width $\Gamma_{T}=(108.4 \pm 15.4)~ \mathrm{MeV} $ of the tensor state $T$ provide useful information for experimental studies of fully heavy four-quark exotic structures., Comment: 12 Pages and 4 Figures

We conduct an investigation on the spin $\frac{3}{2}\rightarrow \frac{1}{2}$ semileptonic weak transition of single heavy baryons for the exclusive decay $\Sigma_{b}^{*0}\,\rightarrow\, \Sigma_{c}^{+}\,\ell\,\bar{\nu}_{\ell}$ in three possible lepton channels within the three point QCD sum rule method. We compute the responsible form factors of this semileptonic decay by incorporating both perturbative and non-perturbative contributions of the operator product expansion series up to mass dimension six. Having acquired the form factors, the decay widths of the processes in all lepton channels are determined. Our findings as well as possible future experimental information can be employed in order to check the SM predictions and explore the possibility of new physics in heavy baryonic decay channels., Comment: 22 Pages, 5 Figures and 7 Tables

Force/torque sensing is an important modality for robotic manipulation, but commodity solutions, generally developed with other applications in mind, do not generally fit the needs of robot hands. This paper introduces a novel method for six-axis force/torque sensing, using LEDs to sense the displacement between two plates connected by a transparent elastomer. Our method allows for finger-size packaging with no amplification electronics, low cost manufacturing, and easy integration into a complete hand. On test forces between 0-2 N, our prototype sensor exhibits a mean error between 0.05 and 0.07 N across the three force directions, suggesting future applicability to fine manipulation tasks.

We investigate the weak semileptonic decay of the $ \Omega^{+}_{ccb} \rightarrow \Xi^{++}_{cc} ~{\ell}\bar\nu_{\ell}$, where a triply heavy baryon with spin 1/2 decays into the observed doubly heavy baryon with spin 1/2, using QCD sum rule method in all lepton channels. We compute the six relevant vector and axial vector form factors entering the low energy matrix elements in full theory. The invariant form factors are building blocks, using the fit faction of which in terms of $ q^2 $ in whole physical region, we calculate the exclusive widths in three lepton channels. Our predictions may help the present and future experiments in the course of their search for doubly and triply heavy baryons., Comment: 20 Pages, 8 Figures and 4 Tables

Mass and width of a hidden charm-bottom axial-vector structure $T$ containing $bc \overline{b}\overline{c}$ quarks are calculated in QCD sum rule framework. It is treated as a diquark-antidiquark state built of scalar diquark and axial-vector antidiquark components. The mass of $T$ is computed using the two-point sum rule method. The width of this particle is evaluated by considering eight decay modes: The decays to $\eta _{b}J/\psi $, $\eta _{c}\Upsilon (1S)$, $B_{c}^{-}B_{c}^{\ast +}$, and $B_{c}^{+}B_{c}^{\ast -}$ are dissociation processes, in which all initial quarks are distributed between the final-state particles. Four decays to $DD$ mesons with appropriate charges and spin-parities are channels generated due to the annihilation of $b$ and $\overline{b}$ quarks from $T$. Partial widths for all of these processes are obtained by employing the three-point sum rule approach necessary to find the strong couplings at relevant tetraquark-meson-meson vertices. Our results for the mass $m=(12715\pm 90)~\mathrm{MeV}$ and width $\Gamma[T] =(104.2 \pm 11.2)~ \mathrm{MeV}$ of the tetraquark $T$, as well as its numerous decay channels explored in this article are useful for ongoing and future experimental investigations of fully heavy resonances., Comment: 10 Pages and 4 Figures

Simulated patient systems play a crucial role in modern medical education and research, providing safe, integrative learning environments and enabling clinical decision-making simulations. Large Language Models (LLM) could advance simulated patient systems by replicating medical conditions and patient-doctor interactions with high fidelity and low cost. However, ensuring the effectiveness and trustworthiness of these systems remains a challenge, as they require a large, diverse, and precise patient knowledgebase, along with a robust and stable knowledge diffusion to users. Here, we developed AIPatient, an advanced simulated patient system with AIPatient Knowledge Graph (AIPatient KG) as the input and the Reasoning Retrieval-Augmented Generation (Reasoning RAG) agentic workflow as the generation backbone. AIPatient KG samples data from Electronic Health Records (EHRs) in the Medical Information Mart for Intensive Care (MIMIC)-III database, producing a clinically diverse and relevant cohort of 1,495 patients with high knowledgebase validity (F1 0.89). Reasoning RAG leverages six LLM powered agents spanning tasks including retrieval, KG query generation, abstraction, checker, rewrite, and summarization. This agentic framework reaches an overall accuracy of 94.15% in EHR-based medical Question Answering (QA), outperforming benchmarks that use either no agent or only partial agent integration. Our system also presents high readability (median Flesch Reading Ease 77.23; median Flesch Kincaid Grade 5.6), robustness (ANOVA F-value 0.6126, p>0.1), and stability (ANOVA F-value 0.782, p>0.1). The promising performance of the AIPatient system highlights its potential to support a wide range of applications, including medical education, model evaluation, and system integration., Comment: 42 pages, 6 figures, 7 tables

In this paper, we propose a framework to enhance the robustness of the neural models by mitigating the effects of process-induced and aging-related variations of analog computing components on the accuracy of the analog neural networks. We model these variations as the noise affecting the precision of the activations and introduce a denoising block inserted between selected layers of a pre-trained model. We demonstrate that training the denoising block significantly increases the model's robustness against various noise levels. To minimize the overhead associated with adding these blocks, we present an exploration algorithm to identify optimal insertion points for the denoising blocks. Additionally, we propose a specialized architecture to efficiently execute the denoising blocks, which can be integrated into mixed-signal accelerators. We evaluate the effectiveness of our approach using Deep Neural Network (DNN) models trained on the ImageNet and CIFAR-10 datasets. The results show that on average, by accepting 2.03% parameter count overhead, the accuracy drop due to the variations reduces from 31.7% to 1.15%.

In this paper, we propose a network architecture where two types of aerial infrastructures together with a ground station provide connectivity to a remote area. A high altitude platform station (HAPS) is equipped with reconfigurable intelligent surface (RIS), so-called HAPS-RIS, to be exploited to assist the unmanned aerial vehicle (UAV)-based wireless networks. A key challenge in such networks is the restricted number of UAVs, which limits full coverage and leaves some users unsupported. To tackle this issue, we propose a hierarchical bilevel optimization framework including a leader and a follower problem. The users served by HAPS-RIS are in a zone called HAPS-RIS zone and the users served by the UAVs are in another zone called UAV zone. In the leader problem, the goal is to establish the zone boundary that maximizes the number of users covered by HAPS-RIS while ensuring that users in this zone meet their rate requirements. This is achieved through an algorithm that integrates RIS clustering, subcarrier allocation, and zone determination. The follower problem focuses on minimizing the number of UAVs required, ensuring that the rate requirements of the users in the UAV zone are met. This is addressed using an algorithm that employs k-means clustering and subcarrier allocation. Our study reveals that increasing the number of RIS elements significantly decreases the number of required UAVs.

Large scale validation and uncertainty quantification are essential in the experimental design, control, and operations of fusion reactors. Reduced models and increasing computational power means that it is possible to run many simulations, yet setting up simulation runs remaining a time-consuming and error-prone process that involves many manual steps. duqtools is an open-source workflow tool written in Python for that addresses this bottleneck by automating the set up of new simulations. This enables uncertainty quantification and large scale validation of fusion energy modelling simulations. In this work, we demonstrate how duqtools can be used to set up and launch 2000 different simulations of plasma experiments to validate aspects of the JINTRAC modelling suite. With this large-scale validation we identified issues in preserving data consistency in model initialization of the current ($I(p)$) distribution. Furthermore, we used duqtools for sensitivity analysis on the QLKNN-jetexp-15D surrogate model to verify its correctness in multiple regimes.

Via numerical simulations and analytical calculations, depletion forces are studied in mixtures of small and big particles that interact via soft repulsive potentials. While big particles are spherical, small particles are nonspherical with shapes that vary gradually, from squares to rods via intermediate shapes. The mixtures are studied for a wide range of densities and temperature. Depletion forces and their resulting potentials depend on the interplay of shape, temperature and density, an argument that is elaborated qualitatively and quantitatively. While in some thermodynamic conditions, depletion potentials of distinct shapes are distinguishable, in different conditions, they are very similar. Finally, I propose novel computational models and experiments for further investigation of the effect of morphology on phase separation in and out of thermal equilibrium.

We study the semileptonic decay of the doubly heavy baryon $ \Xi^{++}_{cc} $ into the singly heavy baryon $ \Xi^+_{c}$ within the three-point QCD sum rule approach in two possible lepton channels. Our analysis includes perturbative as well as nonperturbative condensation contributions up to dimension 5. We evaluate the form factors of this semileptonic decay entering the amplitude described by the vector and axial vector transition currents. The fit functions of the form factors with respect to the transferred momentum squared are utilized to predict the decay widths and branching ratios of the $ \Xi^{++}_{cc}\rightarrow \Xi^+_{c} \bar{\ell}\nu_{\ell}$ channels. We compare our findings with other predictions in the literature. Our outcomes can be useful for experimental groups in their search for the weak decays of doubly heavy baryons and may be checked via future experiments such as LHCb., Comment: 17 Pages, 6 Figures and 4 Tables

Graph Neural Networks (GNNs) excel in handling graph-structured data but often underperform in link prediction tasks compared to classical methods, mainly due to the limitations of the commonly used Message Passing GNNs (MPNNs). Notably, their ability to distinguish non-isomorphic graphs is limited by the 1-dimensional Weisfeiler-Lehman test. Our study presents a novel method to enhance the expressivity of GNNs by embedding induced subgraphs into the graph Laplacian matrix's eigenbasis. We introduce a Learnable Lanczos algorithm with Linear Constraints (LLwLC), proposing two novel subgraph extraction strategies: encoding vertex-deleted subgraphs and applying Neumann eigenvalue constraints. For the former, we conjecture that LLwLC establishes a universal approximator, offering efficient time complexity. The latter focuses on link representations enabling differentiation between $k$-regular graphs and node automorphism, a vital aspect for link prediction tasks. Our approach results in an extremely lightweight architecture, reducing the need for extensive training datasets. Empirically, our method improves performance in challenging link prediction tasks across benchmark datasets, establishing its practical utility and supporting our theoretical findings. Notably, LLwLC achieves 20x and 10x speedup by only requiring 5% and 10% data from the PubMed and OGBL-Vessel datasets while comparing to the state-of-the-art.

The proton-radius puzzle refers to the discrepancy observed in measurements of the proton's charge radius when using different methods. This inconsistency has prompted extensive research and debate within the physics community, as it challenges the understanding of quantum electrodynamics and the fundamental properties of protons. In the present study, we determine the charge and magnetic radii of the proton and neutron through a global analysis of the generalized parton distributions (GPDs) at zero skewness for the first time. Our results challenge the measurements or analyses in which the nucleon's radii are extracted considering just a particular experiment or observable, or using only data points covering particular kinematic regions. We emphasize that simultaneous analysis of all available experimental data related to the radii of the nucleons would be preferable to determine their exact values. The final results obtained from our analysis by this way are: $ r_{pE} = 0.8558 \pm 0.0135~\textrm{fm} $, $ r_{pM} = 0.8268 \pm 0.0533~\textrm{fm} $, $ \left = -0.1181 \pm 0.0270~\textrm{fm}^2 $, and $ r_{nM} = 0.8367 \pm 0.0845~\textrm{fm} $., Comment: 8 Pages, 4 Figures and 2 Tables

This research introduces an innovative method for adaptive traffic signal control (ATSC) through the utilization of multi-objective deep reinforcement learning (DRL) techniques. The proposed approach aims to enhance control strategies at intersections while simultaneously addressing safety, efficiency, and decarbonization objectives. Traditional ATSC methods typically prioritize traffic efficiency and often struggle to adapt to real-time dynamic traffic conditions. To address these challenges, the study suggests a DRL-based ATSC algorithm that incorporates the Dueling Double Deep Q Network (D3QN) framework. The performance of this algorithm is assessed using a simulated intersection in Changsha, China. Notably, the proposed ATSC algorithm surpasses both traditional ATSC and ATSC algorithms focused solely on efficiency optimization by achieving over a 16% reduction in traffic conflicts and a 4% decrease in carbon emissions. Regarding traffic efficiency, waiting time is reduced by 18% compared to traditional ATSC, albeit showing a slight increase (0.64%) compared to the DRL-based ATSC algorithm integrating the D3QN framework. This marginal increase suggests a trade-off between efficiency and other objectives like safety and decarbonization. Additionally, the proposed approach demonstrates superior performance, particularly in scenarios with high traffic demand, across all three objectives. These findings contribute to advancing traffic control systems by offering a practical and effective solution for optimizing signal control strategies in real-world traffic situations.

Helmholtz stereopsis is one the versatile techniques for 3D geometry reconstruction from 2D images of objects with unknown and arbitrary reflectance surfaces. HS eliminates the need for surface reflectance, a challenging parameter to measure, based on the Helmholtz reciprocity principle. Its Bayesian formulation using maximum a posteriori (MAP) probability approach has significantly improved reconstruction accuracy of HS method. This framework enables the inclusion of smoothness priors which enforces observations and neighborhood information in the formulation. We used Markov Random Fields (MRF) which is a powerful tool to integrate diverse prior contextual information and solved the MAP-MRF using belief propagation algorithm. We propose a new smoothness function utilizing the normal field integration method for refined depth estimation within the Bayesian framework. Utilizing three pairs of images with different viewpoints, our approach demonstrates superior depth label accuracy compared to conventional Bayesian methods. Experimental results indicate that our proposed method yields a better depth map with reduced RMS error, showcasing its efficacy in improving depth estimation within Helmholtz stereopsis., Comment: 19 pages

Coupling the fields may lead to the emergence of new phenomena. In the realm of classical fields and nonlinear systems, extensive research has been conducted on their solitary and soliton solutions. In the conducted studies, typically two $\phi^4$ systems, or two sine-Gordon systems, have been coupled. The sine-Gordon system exhibits diverse solutions, all well-behaved, with its soliton solutions fully understood. On the other hand, the $\phi^4$ system, which is significant in field theory, has solitary solutions, but these solutions are not solitonic. For example, from a pair of kink and antikink, we cannot construct a bound state; or that after a collision, these two solutions do not revert to their initial status and become disrupted. In this study, we couple a $\phi^4$ system with a sine-Gordon system to impart stability from the sine-Gordon system to the $\phi^4$ system. We have demonstrated that for a coupled $\phi^4$ and sine-Gordon system, this expectation is somewhat met., Comment: 12 pages, 7 figurs

We predict a near-threshold ("zero energy") peak in multi-photon ionization for a dynamical regime where the photon frequency is large compared to the binding energy of the electron. The peak position does not depend on the laser frequency, but on the binding energy and the pulse duration. The effect originates from the fact that bound-continuum dipole transitions are stronger than continuum-continuum ones. To clearly observe this zero-energy photoelectric effect, the spectral width of the laser pulse should be comparable to the binding energy of the ionized orbital, and the second ionization potential should be larger than the photon energy. This suggests negative ions as ideal candidates for corresponding experiments., Comment: 5 pages, 5 figures, 1 table

Parameters of the heavy four-quark scalar meson $T_{\mathrm{bc\overline{b} \overline{c}}}$ with content $bc \overline{b}\overline{c}$ are calculated by means of the sum rule method. This structure is considered as a diquark-antidiquark state built of scalar diquark and antidiquark components. The mass and current coupling of $T_{\mathrm{bc\overline{b} \overline{c}}}$ are evaluated in the context of the two-point sum rule approach. The full width of this tetraquark is estimated by taking into account two types of its possible strong decay channels. First class includes dissociation of $T_{\mathrm{bc\overline{b}\overline{c}}} $ to mesons $\eta_c\eta_{b}$, $B_{c}^{+}B_{c}^{-}$, $B_{c}^{\ast +}B_{c}^{\ast -}$ and $B_{c}^{+}(1^3P_{0})B_{c}^{\ast-}$. Another type of processes are generated by annihilations $\overline{b}b \to \overline{q}q$ of constituent $ b$-quarks which produces the final-state charmed meson pairs $D^{+}D^{-}$, $ D^{0} \overline{D}^{0}$, $D^{*+}D^{*-}$, and $D^{*0}\overline{D}^{*0}$. Partial width all of these decays are found using the three-point sum rule method which is required to calculate strong couplings at corresponding meson-meson-tetraquark vertices. Predictions obtained for the mass $m=(12697 \pm 90)~\mathrm{MeV}$ and width $\Gamma[T_{\mathrm{bc\overline{b}\overline{c} }}]=(142.4 \pm 16.9)~ \mathrm{MeV}$ of this state are compared with alternative results, and are useful for further experimental investigations of fully heavy resonances., Comment: 11 Pages, 4 Figures and 1 Table

The deployment of Vision Transformers (ViTs) on hardware platforms, specially Field-Programmable Gate Arrays (FPGAs), presents many challenges, which are mainly due to the substantial computational and power requirements of their non-linear functions, notably layer normalization, softmax, and Gaussian Error Linear Unit (GELU). These critical functions pose significant obstacles to efficient hardware implementation due to their complex mathematical operations and the inherent resource count and architectural limitations of FPGAs. PEANO-ViT offers a novel approach to streamlining the implementation of the layer normalization layer by introducing a division-free technique that simultaneously approximates the division and square root function. Additionally, PEANO-ViT provides a multi-scale division strategy to eliminate division operations in the softmax layer, aided by a Pade-based approximation for the exponential function. Finally, PEANO-ViT introduces a piece-wise linear approximation for the GELU function, carefully designed to bypass the computationally intensive operations associated with GELU. In our comprehensive evaluations, PEANO-ViT exhibits minimal accuracy degradation (<= 0.5% for DeiT-B) while significantly enhancing power efficiency, achieving improvements of 1.91x, 1.39x, 8.01x for layer normalization, softmax, and GELU, respectively. This improvement is achieved through substantial reductions in DSP, LUT, and register counts for these non-linear operations. Consequently, PEANO-ViT enables efficient deployment of Vision Transformers on resource- and power-constrained FPGA platforms.

Low-rank adaptation (LoRA) has become the default approach to fine-tune large language models (LLMs) due to its significant reduction in trainable parameters. However, trainable parameter demand for LoRA increases with increasing model embedding dimensions, leading to high compute costs. Additionally, its backward updates require storing high-dimensional intermediate activations and optimizer states, demanding high peak GPU memory. In this paper, we introduce large model fine-tuning via spectrally decomposed low-dimensional adaptation (LaMDA), a novel approach to fine-tuning large language models, which leverages low-dimensional adaptation to achieve significant reductions in trainable parameters and peak GPU memory footprint. LaMDA freezes a first projection matrix (PMA) in the adaptation path while introducing a low-dimensional trainable square matrix, resulting in substantial reductions in trainable parameters and peak GPU memory usage. LaMDA gradually freezes a second projection matrix (PMB) during the early fine-tuning stages, reducing the compute cost associated with weight updates to enhance parameter efficiency further. We also present an enhancement, LaMDA++, incorporating a ``lite-weight" adaptive rank allocation for the LoRA path via normalized spectrum analysis of pre-trained model weights. We evaluate LaMDA/LaMDA++ across various tasks, including natural language understanding with the GLUE benchmark, text summarization, natural language generation, and complex reasoning on different LLMs. Results show that LaMDA matches or surpasses the performance of existing alternatives while requiring up to 17.7x fewer parameter updates and up to 1.32x lower peak GPU memory usage during fine-tuning. Code will be publicly available.

The pseudoscalar and vector four-quark states $bb\overline{c}\overline{c}$ are studied in the context of the QCD sum rule method. We model $T_{\mathrm{ \ PS}} $ and $T_{\mathrm{V}}$ as structures built of diquarks $ b^{T}C\gamma_{5}b$, $\overline{c}C\overline{c}^{T}$ and $b^{T}C\gamma _{5}b$ , $\overline{c}C\gamma_{\mu}\gamma_{5}\overline{c}^{T}$, respectively, with $ C$ being the charge conjugation matrix. The spectroscopic parameters of the tetraquarks $T_{\mathrm{PS}}$ and $T_{\mathrm{V}}$, i.e., their masses and current couplings are calculated using QCD two-point sum rule method. We evaluate the full widths of $T_{\mathrm{PS}}$ and $T_{\mathrm{V}}$ by taking into account their kinematically allowed decay channels. In the case of the pseudoscalar particle they are processes $T_{\mathrm{PS}} \to B_{c}^{-}B_{c}^{\ast -}$, $B_{c}^{-}B_{c}^{-}(1^{3}P_{0})$ and $B_{c}^{\ast -}B_{c}^{-}(1^{1}P_{1})$. The vector state $T_{\mathrm{V}}$ can dissociate to meson pairs $2 B_{c}^{-}$, $2 B_{c}^{\ast -}$ and $ B_{c}^{-}B_{c}^{-}(1^{1}P_{1})$. Partial widths of these decays are determined by the strong couplings at relevant tetraquark-meson-meson vertices, which evaluated in the context of the three-point sum rule approach. Predictions obtained for the mass and full width of the pseudoscalar $m =(13.092\pm 0.095)~\mathrm{GeV}$, $\Gamma _{\mathrm{PS} }=(63.7\pm 13.0)~\mathrm{MeV}$ and vector $\widetilde{m} =(13.15\pm 0.10)~ \mathrm{GeV}$, $\Gamma_{\mathrm{V}}=(53.5\pm 10.3)~\mathrm{MeV}$ tetraquarks can be useful for analyses of different four-quark resonances., Comment: 14 Pages, 6 Figures and 2 Tables

Developing therapeutics is a lengthy and expensive process that requires the satisfaction of many different criteria, and AI models capable of expediting the process would be invaluable. However, the majority of current AI approaches address only a narrowly defined set of tasks, often circumscribed within a particular domain. To bridge this gap, we introduce Tx-LLM, a generalist large language model (LLM) fine-tuned from PaLM-2 which encodes knowledge about diverse therapeutic modalities. Tx-LLM is trained using a collection of 709 datasets that target 66 tasks spanning various stages of the drug discovery pipeline. Using a single set of weights, Tx-LLM simultaneously processes a wide variety of chemical or biological entities(small molecules, proteins, nucleic acids, cell lines, diseases) interleaved with free-text, allowing it to predict a broad range of associated properties, achieving competitive with state-of-the-art (SOTA) performance on 43 out of 66 tasks and exceeding SOTA on 22. Among these, Tx-LLM is particularly powerful and exceeds best-in-class performance on average for tasks combining molecular SMILES representations with text such as cell line names or disease names, likely due to context learned during pretraining. We observe evidence of positive transfer between tasks with diverse drug types (e.g.,tasks involving small molecules and tasks involving proteins), and we study the impact of model size, domain finetuning, and prompting strategies on performance. We believe Tx-LLM represents an important step towards LLMs encoding biochemical knowledge and could have a future role as an end-to-end tool across the drug discovery development pipeline.

Vision-Language Models (VLMs) have recently seen significant advancements through integrating with Large Language Models (LLMs). The VLMs, which process image and text modalities simultaneously, have demonstrated the ability to learn and understand the interaction between images and texts across various multi-modal tasks. Reverse designing, which could be defined as a complex vision-language task, aims to predict the edits and their parameters, given a source image, an edited version, and an optional high-level textual edit description. This task requires VLMs to comprehend the interplay between the source image, the edited version, and the optional textual context simultaneously, going beyond traditional vision-language tasks. In this paper, we extend and fine-tune MiniGPT-4 for the reverse designing task. Our experiments demonstrate the extensibility of off-the-shelf VLMs, specifically MiniGPT-4, for more complex tasks such as reverse designing. Code is available at this \href{https://github.com/VahidAz/MiniGPT-Reverse-Designing}, Comment: 8 pages, 7 figures

This work introduces the Supervised Expectation-Maximization Framework (SEMF), a versatile and model-agnostic approach for generating prediction intervals in datasets with complete or missing data. SEMF extends the Expectation-Maximization algorithm, traditionally used in unsupervised learning, to a supervised context, leveraging latent variable modeling for uncertainty estimation. Extensive empirical evaluations across 11 tabular datasets show that SEMF often achieves narrower normalized prediction intervals and higher coverage rates than traditional quantile regression methods. Furthermore, SEMF can be integrated with machine learning models like gradient-boosted trees and neural networks, highlighting its practical applicability. The results indicate that SEMF enhances uncertainty quantification, particularly in scenarios with complete data.

Understanding human behavior fundamentally relies on accurate 3D human pose estimation. Graph Convolutional Networks (GCNs) have recently shown promising advancements, delivering state-of-the-art performance with rather lightweight architectures. In the context of graph-structured data, leveraging the eigenvectors of the graph Laplacian matrix for positional encoding is effective. Yet, the approach does not specify how to handle scenarios where edges in the input graph are missing. To this end, we propose a novel positional encoding technique, PerturbPE, that extracts consistent and regular components from the eigenbasis. Our method involves applying multiple perturbations and taking their average to extract the consistent and regular component from the eigenbasis. PerturbPE leverages the Rayleigh-Schrodinger Perturbation Theorem (RSPT) for calculating the perturbed eigenvectors. Employing this labeling technique enhances the robustness and generalizability of the model. Our results support our theoretical findings, e.g. our experimental analysis observed a performance enhancement of up to $12\%$ on the Human3.6M dataset in instances where occlusion resulted in the absence of one edge. Furthermore, our novel approach significantly enhances performance in scenarios where two edges are missing, setting a new benchmark for state-of-the-art.

The twisted kagome family comprises a spectrum of configurations, that can be realized through the sweep of a single configurational degree of freedom known as twist angle. Recently, it has been shown that certain pairs of configurations along this sweep are linked by duality transformations and display matching phonon spectra. In this work, we introduce an inter-cell connection system that spreads the lattice in the dimension orthogonal to the tessellation plane. The resulting 3D character of the lattice allows us to sweep the entirety of the twist angle spectrum, including all the compact configurations featuring overlapping triangles that, in a strictly 2D space, are forbidden. Duality provides precious guidance in interpreting the availability of floppy mechanisms arising in the compact configurations through the one-to-one correspondence with their expanded counterparts. Our focus is on the compact configuration corresponding to a null twist angle, where the lattice degenerates to a chain. From the perspective of the chain, several of the local connections between neighboring lattice cells play the role of non-local long-range interactions between cells of the chain. We demonstrate experimentally some peculiar behavior that results from such non-locality, including a selective activation of floppy sequences that is informed by the direction of loading., Comment: 8 pages, 10 figures, and 4 ancillary videos