Current Vision-Language Models (VLMs) fail at quantitative spatial reasoning because their architectures destroy pixel-level information required for counting and measurements. Vision encoders compress images through patch embeddings, reducing spatial indexing and losing the precise pixel-level tracking required for accurate counting. We present two contributions to address this fundamental limitation. First, we introduce SQuID (Satellite Quantitative Intelligence Dataset), a benchmark of 2,000 satellite image Question-Answer pairs with both numerical range and categorical answers, designed to evaluate quantitative spatial reasoning. The dataset spans three difficulty tiers with annotations automatically generated from human labels and their learned variability. Second, we propose QVLM (Quantitative Vision-Language Model), a code-generation architecture that maintains pixel precision by decoupling language understanding from visual analysis. Instead of encoding images into embeddings, QVLM generates executable code that first calls a segmentation model to obtain pixel-level masks, then operates directly on these masks, preserving spatial indexing throughout the reasoning process. O
The incidence of advanced renal cell carcinoma(RCC) has been rising, presenting significant challenges due to the limited efficacy and severe side effects of traditional radiotherapy and chemotherapy. While combination immunotherapies show promise, optimizing treatment strategies remains difficult due to individual heterogeneity. To address this, we developed a Quantitative Cancer-Immunity Cycle (QCIC) model that integrates ordinary differential equations with stochastic modelling to quantitatively characterize and predict tumor evolution in patients with advanced RCC. By systematically integrating quantitative systems pharmacology principles with biological mechanistic knowledge, we constructed a virtual patient cohort and calibrated the model parameters using clinical immunohistochemistry data to ensure biological validity. To enhance predictive performance, we coupled the model with pharmacokinetic equations and defined the Tumor Response Index (TRI) as a quantitative metric of efficacy. Systematic analysis of the QCIC model allowed us to determine an optimal treatment regimen for the combination of bevacizumab and atezolizumab and identify tumor biomarkers with clinical predict
As neural components are increasingly embedded in existing symbolic software -- including safety-critical systems -- the question arises of how to specify and enforce the safety of the newly introduced neural parts. Unlike traditional logical specifications, these must be amenable not only to the standard Boolean interpretation, but also to training and optimisation. The latter calls for a quantitative interpretation of the logical syntax, subject to further requirements such as smoothness and differentiability. Moreover, the qualitative and quantitative sides of the logic must share a unifying proof-theoretic and categorical semantics. Finally, the new logic should link cleanly to the substructural and program logics that underpin the verification of existing symbolic programs. In this paper, we present a logic that ticks all of these boxes. We introduce a family of calculi, pQLL, indexed by a hardness degree $p$, prove a cut-elimination theorem for them, and establish completeness with respect to enriched residuated `soft' lattices. At $p = \infty$, \pQLL reduces to multiplicative additive linear logic (MALL), and provability in pQLL converges to provability in MALL as $p \to \in
The search for extraterrestrial life in the Solar System and beyond is a key science driver in astrobiology, planetary science, and astrophysics. A critical step is the identification and characterization of potential habitats, both to guide the search and to interpret its results. However, a well-accepted, self-consistent, flexible, and quantitative terminology and method of assessment of habitability are lacking. Our paper fills this gap based on a three year-long study by the NExSS Quantitative Habitability Science Working Group. We reviewed past studies of habitability, but find that the lack of a universally valid definition of life prohibits a universally applicable definition of habitability. A more nuanced approach is needed. We introduce a quantitative habitability assessment framework (QHF) that enables self-consistent, probabilistic assessment of the compatibility of two models: First, a habitat model, which describes the probability distributions of key conditions in the habitat. Second, a viability model, which describes the probability that a metabolism is viable given a set of environmental conditions. We provide an open-source implementation of this framework and fo
We investigate the convergence of symmetric stochastic differential games with interactions via control, where the volatility terms of both idiosyncratic and common noises are controlled. We apply the stochastic maximum principle, following the approach of Laurière and Tangpi, to reduce the convergence analysis to the study of forward-backward propagation of chaos. Under the standard monotonicity conditions, we derive quantitative convergence rates for open-loop Nash equilibria of $N$-player stochastic differential games toward the corresponding mean field equilibrium. As a prerequisite, we also establish the well-posedness of the conditional McKean--Vlasov forward-backward stochastic differential equations by the method of continuation. Moreover, we analyze a specific class of linear-quadratic settings to demonstrate the applicability of our main result.
Quantitative investment (quant) is an emerging, technology-driven approach in asset management, increasingy shaped by advancements in artificial intelligence. Recent advances in deep learning and large language models (LLMs) for quant finance have improved predictive modeling and enabled agent-based automation, suggesting a potential paradigm shift in this field. In this survey, taking alpha strategy as a representative example, we explore how AI contributes to the quantitative investment pipeline. We first examine the early stage of quant research, centered on human-crafted features and traditional statistical models with an established alpha pipeline. We then discuss the rise of deep learning, which enabled scalable modeling across the entire pipeline from data processing to order execution. Building on this, we highlight the emerging role of LLMs in extending AI beyond prediction, empowering autonomous agents to process unstructured data, generate alphas, and support self-iterative workflows.
This study compares two statistical approaches to image reconstruction in single-photon emission computed tomography (SPECT). We evaluated the widely used Ordered Subset Expectation Maximization (OSEM) algorithm and the newer Maximum a Posteriori approach with Entropy prior (MAP-Ent) approach in the context of quantifying radiopharmaceutical uptake in pathological lesions. Numerical experiments were performed using a digital twin of the standardized NEMA IEC phantom, which contains six spheres of varying diameters to simulate lesions. Quantitative accuracy was assessed using the maximum recovery coefficient (RCmax), defined as the ratio of the reconstructed maximum activity to the true value. The study shows that OSEM exhibits unstable convergence during iterations, leading to noise and edge artifacts in lesion images. Post-filtering stabilizes the reconstruction and ensures convergence, producing RCmax-size curves that could be used as correction factors in clinical evaluations. However, this approach significantly underestimates uptake in small lesions and may even lead to the complete loss of small lesions on reconstructed images. In contrast, MAP-Ent demonstrates fundamentally
Reinforcement learning (RL) is a subfield of machine learning that has been used in many fields, such as robotics, gaming, and autonomous systems. There has been growing interest in using RL for quantitative trading, where the goal is to make trades that generate profits in financial markets. This paper presents the use of RL for quantitative trading and reports a case study based on an RL-based trading algorithm. The results show that RL can be a useful tool for quantitative trading and can perform better than traditional trading algorithms. The use of reinforcement learning for quantitative trading is a promising area of research that can help develop more sophisticated and efficient trading systems. Future research can explore the use of other reinforcement learning techniques, the use of other data sources, and the testing of the system on a range of asset classes. Together, our work shows the potential in the use of reinforcement learning for quantitative trading and the need for further research and development in this area. By developing the sophistication and efficiency of trading systems, it may be possible to make financial markets more efficient and generate higher retur
Large-scale AI models such as GPT-4 have accelerated the deployment of artificial intelligence across critical domains including law, healthcare, and finance, raising urgent questions about trust and transparency. This study investigates the relationship between explainability and user trust in AI systems through a quantitative experimental design. Using an interactive, web-based loan approval simulation, we compare how different types of explanations, ranging from basic feature importance to interactive counterfactuals influence perceived trust. Results suggest that interactivity enhances both user engagement and confidence, and that the clarity and relevance of explanations are key determinants of trust. These findings contribute empirical evidence to the growing field of human-centered explainable AI, highlighting measurable effects of explainability design on user perception
We show that the spatial increments of the KPZ fixed point starting from arbitrary initial data, exhibit strong quantitative comparison against rate two Brownian motion on compacts. The above estimates are uniform in the initial data supported in some compact set. As applications, we obtain a one-sided large deviation inequality for spatial increments of the KPZ fixed point and show that the Wiener density of the centred KPZ fixed point started from arbitrary initial data has finite entropy.
Quantitative investment aims to maximize the return and minimize the risk in a sequential trading period over a set of financial instruments. Recently, inspired by rapid development and great potential of AI technologies in generating remarkable innovation in quantitative investment, there has been increasing adoption of AI-driven workflow for quantitative research and practical investment. In the meantime of enriching the quantitative investment methodology, AI technologies have raised new challenges to the quantitative investment system. Particularly, the new learning paradigms for quantitative investment call for an infrastructure upgrade to accommodate the renovated workflow; moreover, the data-driven nature of AI technologies indeed indicates a requirement of the infrastructure with more powerful performance; additionally, there exist some unique challenges for applying AI technologies to solve different tasks in the financial scenarios. To address these challenges and bridge the gap between AI technologies and quantitative investment, we design and develop Qlib that aims to realize the potential, empower the research, and create the value of AI technologies in quantitative in
An emergent challenge in geriatric care is improving the quality of care, which requires insight from stakeholders. Qualitative methods offer detailed insights, but they can be biased and have limited generalizability, while quantitative methods may miss nuances. Network-based approaches, such as quantitative ethnography (QE), can bridge this methodological gap. By leveraging the strengths of both methods, QE provides profound insights into need-finding interviews. In this paper, to better understand geriatric care attitudes, we interviewed ten nursing assistants, used QE to analyze the data, and compared their daily activities in real life with training experiences. A two-sample t-test with a large effect size (Cohen's d=1.63) indicated a significant difference between real-life and training activities. The findings suggested incorporating more empathetic training scenarios into the future design of our geriatric care simulation. The results have implications for human-computer interaction and human factors. This is illustrated by presenting an example of using QE to analyze expert interviews with nursing assistants as caregivers to inform subsequent design processes.
We obtain a solution to a bordism version of Gromov's linearity problem over a large family of acyclic groups, for manifolds with arbitrary dimension. Every group embeds into some acyclic group in this family. Thus, the linear bordism problem has an affirmative solution over a possibly enlarged acyclic group. Our result holds in both PL and smooth categories, and for both oriented and unoriented cases. In the PL case, our results hold without assuming bounded local geometry. As an application, we prove that there is a universal linear bound for the Cheeger-Gromov $L^2$ $ρ$-invariants of PL $(4k-1)$-manifolds associated with arbitrary regular covers. We also show that the minimum number of simplices in a PL triangulation of $(4k-1)$-manifolds with a fixed simple homotopy type is unbounded if the fundamental group has nontrivial torsion. The proof of our main results builds on quantitative algebraic and geometric techniques over the simplicial classifying spaces of groups.
Bacterial sensor systems can be used for the detection and measurement of molecular signal concentrations. The dynamics of the sensor directly depend on the biological properties of the bacterial sensor cells; manipulation of these features in the wet lab enables the engineering and optimization of the bacterial sensor kinetics. This necessitates the development of biologically meaningful computational models for bacterial sensors comprising a variety of different molecular mechanisms, which further facilitates a systematic and quantitative evaluation of optimization strategies. In this work, we dissect the detection chain of bacterial sensors from a mathematical perspective from which we derive, supported by wet-lab data, a complete computational model for a Streptococcus mutans-based bacterial sensor as a case example. We address the engineering of bacterial sensors by investigating the impact of altered bacterial cell properties on the sensor response characteristics, specifically sensor sensitivity and response signal intensity. This is achieved through a sensitivity analysis targeting both the steady-state and transient sensor response characteristics. Alongside the demonstrat
In this paper, we propose a series of fuzzy temporal protoforms in the framework of the automatic generation of quantitative and qualitative natural language descriptions of processes. The model includes temporal and causal information from processes and attributes, quantifies attributes in time during the process life-span and recalls causal relations and temporal distances between events, among other features. Through integrating process mining techniques and fuzzy sets within the usual Data-to-Text architecture, our framework is able to extract relevant quantitative temporal as well as structural information from a process and describe it in natural language involving uncertain terms. A real use-case in the cardiology domain is presented, showing the potential of our model for providing natural language explanations addressed to domain experts.
In this paper, we develop a quantitative K-theory for filtered C*-algebras. Particularly interesting examples of filtered C*-algebras include group C*-algebras, crossed product C*-algebras and Roe algebras. We prove a quantitative version of the six term exact sequence and a quantitative Bott periodicity. We apply the quantitative K-theory to formulate a quantitative version of the Baum-Connes conjecture and prove that the quantitative Baum-Connes conjecture holds for a large class of groups.
What is the demand elasticity of statistical arbitrageurs that invest according to the advice of modern cross-sectional asset pricing models? Thirteen models from the literature exhibit strikingly inelastic demand, in contrast to classical models that rely on statistical arbitrageurs to create elastic market demand for assets. This inelasticity arises from the difficulty of trading against price changes. A quantitative equilibrium model shows that aggregate demand remains inelastic even with these statistical arbitrageurs in the market.
The importance of research teams' diversity for the progress of science is highlighted extensively. Despite the seemingly hegemonic role of hypothesis testing in modern quantitative research, little attention has been devoted to the diversity of quantitative methods, epitomized by the linear model framework of analysis. Using bibliometric data from the Web of Science, we conduct a large-scale and cross-disciplinary assessment of the prevalence of linear-model-based research from 1990 to 2022. In absolute terms, linear models are widely used across all fields of science. In relative terms, three patterns suggest linear models are hegemonic among Social Sciences. First, there is a high and growing prevalence of linear-model-based research. Second, global patterns of linear-model-based research prevalence align with global inequalities in knowledge production. Third, there was a citation premium to linear-model-based research until 2012 for articles' number of citations and for the entire period in terms of having at least one citation. Previous research suggests that the confluence of these patterns may be detrimental to the Social Sciences as it potentially marginalizes theories inc
This paper is concerned with quantitative estimates for the Navier-Stokes equations. First we investigate the relation of quantitative bounds to the behaviour of critical norms near a potential singularity with Type I bound $\|u\|_{L^{\infty}_{t}L^{3,\infty}_{x}}\leq M$. Namely, we show that if $T^*$ is a first blow-up time and $(0,T^*)$ is a singular point then $$\|u(\cdot,t)\|_{L^{3}(B_{0}(R))}\geq C(M)\log\Big(\frac{1}{T^*-t}\Big),\,\,\,\,\,\,R=O((T^*-t)^{\frac{1}{2}-}).$$ We demonstrate that this potential blow-up rate is optimal for a certain class of potential non-zero backward discretely self-similar solutions. Second, we quantify the result of Seregin (2012), which says that if $u$ is a smooth finite-energy solution to the Navier-Stokes equations on $\mathbb{R}^3\times (0,1)$ with $$\sup_{n}\|u(\cdot,t_{(n)})\|_{L^{3}(\mathbb{R}^3)}<\infty\,\,\,\textrm{and}\,\,\,t_{(n)}\uparrow 1,$$ then $u$ does not blow-up at $t=1$. To prove our results we develop a new strategy for proving quantitative bounds for the Navier-Stokes equations. This hinges on local-in-space smoothing results (near the initial time) established by Jia and Šverák (2014), together with quantitative argument
A quantitative measure of stability in stochastic dynamics starts to emerge in recent experiments on bioswitches. This quantity, similar to the potential function in mathematics, is deeply rooted in biology, dated back at the beginning of quantitative description of biological processes: the adaptive landscape of Wright (1932) and the development landscape of Waddington (1940). Nevertheless, its quantitative implication has been frequently challenged by biologists. Recent progresses in quantitative biology begin to meet those outstanding challenges.