150813 research outputs found
Sort by
A Multitask Deep Learning Framework for Clinical Decision-Making in Assisted Reproductive Technology
No full textFertility treatment requires navigating a series of complex and connected decisions, each with tradeoffs that impact both patient outcomes and clinical workload. In in vitro fertilization (IVF), choices about stimulation protocols, embryo transfer count, and treatment timing are interdependent, and each cycle carries uncertainty. Despite increasing data availability, most clinical tools still address these decisions in isolation. As a result, care remains variable across providers, and patients often face unclear guidance at critical moments in the process. To address this gap, we present an integrative deep learning framework that simultaneously optimizes three critical IVF decision points using data from 33,000 treatment cycles (2021- 2023): (1) embryo transfer count recommendation (MAE = 0.716), (2) ovarian stimulation protocol selection (AUC = 0.85), and (3) biochemical pregnancy prediction (AUC = 0.75). Our comprehensive benchmarking pipeline evaluates classical statistical models, ensemble methods (XGBoost), and novel architectures, including TabPFN, an attention-based probabilistic model that achieved comparable performance to top-performing baselines. To enhance clinical trust, we apply SHapley Additive exPlanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME), which consistently uncover biologically meaningful variables, including biomarkers such as Anti-Müllerian Hormone (AMH), patient age, and the day-3 high-quality embryo rate. The final architecture combines: (i) a shared representation learning backbone, (ii) task-specific prediction heads, and (iii) homoscedastic uncertainty weighting for dynamic task balancing. This multitask approach achieves near-state-of-the-art performance (protocol AUC = 0.79, pregnancy AUC = 0.74) while offering practical deployment advantages such as faster inference and fewer parameters than equivalent single-task ensembles. This system’s efficiency, interpretability, and scalability will reshape current IVF practice if adopted in clinical settings, offering a more consistent and data-driven foundation for treatment decisions.MN
First-principles study of SiO 2 / MoS 2 and SiO 2 / WS 2 interfaces: A comparative analysis of surface terminations, van der Waals corrections, and functionals
No full textThis study presents a first-principles investigation of SiO 2 / MoS 2 and SiO 2 / WS 2 interfaces, examining how surface terminations, van der Waals (vdW) corrections, and functional choices impact structural stability and electronic properties. Using density functional theory with generalized gradient approximation (GGA; PBE, PBEsol, revPBE), meta-GGA (SCAN, r 2 SCAN), and hybrid (PBE0) functionals, we assess the effect of vdW correction schemes (D2, D3, Tkatchenko-Scheffler) on interfacial energetics and separation. The results show that vdW corrections are essential for accurate GGA descriptions, while meta-GGAs yield similar accuracy even without them, enabling efficient modeling of SiO 2 /2D heterostructures. Additionally, SiO 2 surface morphology plays a significant role, with fully saturated interfaces showing lower energy and greater interlayer separations. In both SiO 2 / MoS 2 and SiO 2 / WS 2 systems, band gap predictions using PBE0 closely match the experimental values, underscoring the value of hybrid functionals for accurate electronic structure calculations
Search for pair production of heavy particles decaying to a top quark and a gluon in the lepton+jets final state in proton–proton collisions at √s = 13 Te V
No full textA search is presented for the pair production of new heavy resonances, each decaying into a top quark (t) or antiquark and a gluon (g). The analysis uses data recorded with the CMS detector from proton–proton collisions at a center-of-mass energy of 13 Te V at the LHC, corresponding to an integrated luminosity of 138 fb - 1 . Events with one muon or electron, multiple jets, and missing transverse momentum are selected. After using a deep neural network to enrich the data sample with signal-like events, distributions in the scalar sum of the transverse momenta of all reconstructed objects are analyzed in the search for a signal. No significant deviations from the standard model prediction are found. Upper limits at 95% confidence level are set on the product of cross section and branching fraction squared for the pair production of excited top quarks in the t ∗ → tg decay channel. The upper limits range from 120 to 0.8 fb for a t ∗ with spin-1/2 and from 15 to 1.0 fb for a t ∗ with spin-3/2. These correspond to mass exclusion limits up to 1050 and 1700 Ge V for spin-1/2 and spin-3/2 t ∗ particles, respectively. These are the most stringent limits to date on the existence of t ∗ → tg resonances
Vigilis: Leveraging Language Models for Fraud Detection in Mobile Communications and Financial Transactions
No full textAlthough advances in security have strengthened defenses in digital financial systems, attackers increasingly rely on social engineering to achieve their goals. These attacks are difficult to detect and prevent with existing security measures. To address this, we propose Vigilis, a fraud-protected application that employs advanced language models to counter such attacks in calls, texts, and payments. We first collect and make available a corpus of fraudulent calls from the Internet and train lightweight transformer-based models that achieve fraud detection accuracies of up to 94% and 87% on transcript and audio modalities, respectively. We integrate these models into a real-time call system within Vigilis that operates entirely on-device, enabling accurate fraud detection in an efficient and privacy-preserving manner. We then extend Vigilis to incorporate context-aware transaction authentication, where the underlying social context behind a transaction is determined from calls, texts, and browsing history and used to infer the transaction’s validity. By uniquely incorporating social concepts into traditional cybersecurity techniques, we attempt to counter and mitigate issues related to social engineering attacks in financial fraud.M.Eng
Urban Planning for Health Equity Must Employ an Intersectionality Framework
No full textUrban planning for health equity should be guided by an intersectional approach. Intersectionality is an essential framework for understanding the multiple overlapping factors, such as social and economic inequalities, that produce health disparities. We offer four strategies that planning researchers and practitioners can use to develop and integrate an intersectional approach into planning for health equity: challenging implicit and explicit assumptions, building cross-sectoral coalitions united by a shared vision for social and environmental justice, applying transdisciplinary and co-designing approaches throughout the planning process, and using existing tools to evaluate the impact of programs and policies on advancing health equity
Dynamic Scene Editing via Semantically Trained 3D Guassians
No full textImage-based 3D scene reconstruction continues to be a challenge as it involves solving both the sufficient 3D representation problem and the 3D reconstruction itself. One approach to tackle the rendering problem is 3D Gaussian Splatting because of its potential to produce fast and realistic renders via 3D Gaussian representation. With many applications in the entertainment industry, there is motivation in using 3D Gaussian Splatting for not only reconstructing 3D dynamic scenes but also editing them. However, extending the problem to dynamic 3D scenes proves to be a challenging task as it involves discerning the correct representation of a 3D scene while maintaining the capability to render in real time. State-ofthe-art methods have proposed methods that reconstruct dynamic scenes or edit static scenes, but the problem of editing dynamic scenes is still underexplored. This thesis analyzes the feasibility of editing semantically trained Gaussians for dynamic 3D scene editing. By training 3D Gaussians to represent the semantics across the time steps of a dynamic 3D scene, these primitives can be combined with an image editing pipeline to perform real-time, realistic 3D scene editing. Results show that editing segmented 3D Gaussians produces higher-quality and efficient renders as compared to editing without segmentation. However, when evaluated for mainstream applications, results show the impracticality of this pipeline and draw focus to memory and editing limitations that need to be further researched for future advances in 3D Gaussian Splatting.M.Eng
Neo-Panamax Decarbonization via Microreactor Propulsion Conversion
No full textThis study presents a comprehensive feasibility assessment for retrofitting a Neo-Panamax (NPX)
container vessel with nuclear microreactor propulsion to contribute to decarbonization of commercial shipping.
The project selected a 12,000 TEU container vessel as a baseline hull and replaced its WinGD 7x92-B diesel
engine and auxiliary generators with two MIT-designed Organically Cooled Reactors (OCRs), each paired with
a 27MW Mitsubishi steam turbine generator and a Leonardo DRS 36.5MW direct-drive electric motor. Detailed
Computer-Aided Design (CAD) modeling and Finite Element Analysis (FEA) were used to validate seakeeping
performance, optimize system arrangements, and verify the structural integrity of deck reinforcements under
static and buckling loads. Stability and damaged-condition survivability were evaluated using MAXSURF,
demonstrating intact and damaged American Bureau of Shipping (ABS) compliance across operational load
cases. Seakeeping analyses at sea states 4–9 confirmed that motions remain within recoverable righting-arm
limits. A bottom-up financial analysis compared lifecycle costs over 25 years, showing that the retrofit’s 946M projected lifecycle cost of a conventional NPX and yields $405–806M in net savings
when accounting for impending carbon taxes. Key regulatory challenges including absence of propulsion-
specific nuclear regulations and port-entry protocols were identified as primary non-technical hurdles, with
emerging frameworks from industry consortia offering pathways to implementation. Nuclear microreactor
retrofits can be technically and economically viable for large commercial vessels, positioning them as a potent
strategy to meet International Maritime Organization’s (IMO) net-zero targets by 2050
A Dual-Branch Coupled Fourier Neural Operator for High-Resolution Multi-Phase Flow Modeling in Porous Media
No full textThis paper investigates a physics-informed surrogate modeling framework for multi-phase flow in porous media based on the Fourier Neural Operator. Traditional numerical simulators, though accurate, suffer from severe computational bottlenecks due to fine-grid discretizations and the iterative solution of highly nonlinear partial differential equations. By parameterizing the kernel integral directly in Fourier space, the operator provides a discretization-invariant mapping between function spaces, enabling efficient spectral convolutions. We introduce a Dual-Branch Adaptive Fourier Neural Operator with a shared Fourier encoder and two decoders: a saturation branch that uses an inverse Fourier transform followed by a multilayer perceptron and a pressure branch that uses a convolutional decoder. Temporal information is injected via Time2Vec embeddings and a causal temporal transformer, conditioning each forward pass on step index and time step to maintain consistent dynamics across horizons. Physics-informed losses couple data fidelity with residuals from mass conservation and Darcy pressure, enforcing the governing constraints in Fourier space; truncated spectral kernels promote generalization across meshes without retraining. On SPE10-style heterogeneities, the model shifts the infinity-norm error mass into the 10−2 to 10−1 band during early transients and sustains lower errors during pseudo-steady state. In zero-shot three-dimensional coarse-to-fine upscaling from 30 ×110 ×5 to 60 ×220 ×5, it attains 2 =0.90, RMSE = 4.4 ×10−2, and MAE = 3.2 ×10−2, with more than 90% of voxels below five percent absolute error across five unseen layers, while the end-to-end pipeline runs about three times faster than a full-order fine-grid solve and preserves water-flood fronts and channel connectivity. Benchmarking against established baselines indicates a scalable, high-fidelity alternative for high-resolution multi-phase flow simulation in porous media
A Changing Climate Beneath Our Feet: How plant and microbial life in tropical soils are shifting and what that could mean for the future of our warming planet
No full textDiscussions about climate change and carbon sequestration have largely revolved around plant structures we can easily see, like leaves that absorb CO₂ for photosynthesis and woody trunks that store carbon as biomass. Carbon credits that companies and consumers buy to compensate for emissions they’ve produced are primarily calculated based on these parts, as are models that predict climate change impacts. But researchers are now beginning to understand that what we see aboveground is only part of the equation. The other part lies beneath our feet in an intricate, expansive, covert realm where plant roots, microbial communities and soil dynamics interact. These belowground systems are crucial for cycling carbon through the Earth and regulating the climate, but relatively little is known about them compared to aboveground systems. This is especially true in tropical regions, where one-third of the world’s terrestrial carbon storage lies. However, these systems are evolving quickly with climate change, contradicting what models have previously projected. With so many global decisions based on such models, these uncertainties hold planetary significance for our future. A group of scientists is climbing an uphill battle, racing against time to understand this understudied field.S.M
Study on molecular orientation and stratification in RNA-lipid nanoparticles by cryogenic orbitrap secondary ion mass spectrometry
No full textLipid nanoparticle RNA (LNP-RNA) formulations are used for the delivery of vaccines and other therapies. RNA molecules are encapsulated within their interior through electrostatic interactions with positively charged lipids. The identity of the lipids that present at their surface play a role in how they interact with and are perceived by the body and their resultant potency. Here, we use a model formulation to develop cryogenic sample preparation for molecular depth profiling Orbitrap secondary ion mass spectrometry (Cryo-OrbiSIMS) preceded by morphological characterisation using cryogenic transmission electron microscopy (Cryo-TEM). It is found that the depth distribution of individual lipid components is revealed relative to the surface and the RNA cargo defining the core. A preferential lipid orientation can be determined for the 1,2-Dimyristoyl-glycero-3-methox-polyethylene glycol 2000 (DMG-PEG2k) molecule, by comparing the profiles of PEG to DMG fragments. PEG fragments are found immediately during analysis of the LNP surface, while the DMG fragments are deeper, coincident with RNA ions located in the core, in agreement with established models of LNPs. This laboratory-based de novo analysis technique requires no labelling, providing advantages over large facility neutron scattering characterisation