arXiv subjects

Arnab Barua

Publications and source records attributed to Arnab Barua.

8 recordsLinked to original sources

arXiv2026

QuantFlow: A Federated Mamba-Based Post-Transformer Foundation Model for Time-Series Forecasting

Time-series forecasting supports decisions in finance, en-ergy, transportation, public health, and industrial monitoring. Recent foundation models improve transfer across forecast-ing tasks, but many depend on centralized data and Trans-former attention, which restricts their use for long, high-di-mensional, and privacy-sensitive signals. This paper presents QuantFlow, a probabilistic forecasting framework that com-bines inverted sequence embedding, bidirectional Mamba state-space decoders, quantile regression, and federated learning. Each variable is embedded over the complete ob-servation window, processed in forward and reverse direc-tions, and projected to five conditional quantiles. TSMixup expands temporal diversity through Dirichlet-weighted inter-polation while preserving sequence structure. Experiments cover cryptocurrency, traffic, electricity, Electricity Trans-former Temperature, influenza, and weather data. QuantFlow obtains mean squared errors of 0.2834 on ETTm1 and 0.2218 on Weather, and a 20-client non-IID deployment retains use-ful accuracy after three communication rounds without cen-tralizing raw records. The results indicate that selective state-space modelling is a promising basis for scalable, uncer-tainty-aware, and privacy-conscious time-series prediction, while also revealing limitations on irregular epidemiological signals and long-horizon generalization.

cs.LG↗

arXiv2026

Density-dependent growth emerges from Bayesian adaptation of phenotype

Classical models often describe early tumor expansion as exponential growth, yet experimental and clinical evidence shows that tumor populations can deviate systematically from this behavior, exhibiting density dependent proliferation, cooperative low-density growth, intermediate growth optima, and finite upper growth bounds before resource limitation or spatial crowding dominate. These observations raise a common question: why should the per capita growth rate depend on population size? Here, we propose that sensing mismatch provides a mesoscopic link between environmental change and density dependent proliferation. We model the cell as a Bayesian adaptive agent whose coarse grained phenotype evolves on an intrinsic regulatory landscape, while environmental sensing reweights phenotypic states according to how well they account for the extracellular signal statistics generated by the population. In the weak phenotype signal correlation regime, the stationary phenotype distribution is Gaussian, with its mean displaced from the proliferative optimum by a population size-dependent baseline information mismatch. This displacement produces a quadratic penalty in the per capita growth rate. Coupling the framework to a receptor ligand decoding model, we show that basal readout error and nonlinear receptor saturation make the mismatch nonmonotonic in population size. This single structure gives rise to an intermediate proliferation optimum, an Allee survival threshold, a tissue specific capacity, and superlinear scaling at low density. A phase diagram in the phenotype signal coupling and readout-error plane partitions growth into regulated, uncontrolled, and arrested regimes. Thus, density dependent proliferation need not be imposed phenomenologically, but can emerge from cellular sensing and inference.

physics.bio-ph↗

arXiv2025

Geodesic learning

Learning is a fundamental characteristic of living systems, enabling them to comprehend their environments and make informed decisions. These decision-making processes are inherently influenced by available information about their surroundings and specific objectives. There is an intriguing perspective is that each process is highly efficient under a given set of conditions. A key question, then, is how close to optimality it is or how efficient it is under given conditions. Here, the concept of "geodesic learning" as the optimal reference process, with which each process can be compared, is introduced and formulated on the basis of geometry. The probability distribution describing the state of the composite system consisting of the environment, termed the "information bath", and a decision-maker is characterized by use of the entropic quantities. This enables one to study the system in analogy with thermodynamics. Learning processes are expressed as the changes of parameters contained in the distribution. For a geometric interpretation of the processes, the manifold endowed with the Fisher-Rao metric as the Riemannian metric is considered. This framework allows one to conceptualize the optimality of each process as a state change along a geodesic curve on the manifold, which gives rise to geodesic learning. Then, the bivariate Gaussian model is presented, and the processes of geodesic learning and adaptation are analyzed for illustrating this approach.

cond-mat.dis-nn↗

arXiv2024

Navigating the nexus: a perspective of centrosome -cytoskeleton interactions

A structural relationship between the centrosome and cytoskeleton has been recognized for many years. Centrosomes typically reside near the nucleus, establishing and maintaining the nucleus-centrosome axis. This spatial arrangement is critical for determining cell polarity during interphase and ensuring the proper assembly of the spindle apparatus during mitosis. Centrosomes also engage in physical interactions with various components of the cytoskeleton, balancing internal cellular architecture and polarity in a manner specific to tissue type and developmental stage. Numerous crosslinking proteins facilitate these interactions, promoting both cytoskeletal and centrosomal nucleation. This article provides an overview of how cytoskeletal elements and centrosomes coordinate their actions to regulate complex cellular functions such as cell migration, adhesion, and division. The reciprocal influence between cytoskeletal dynamics and centrosomal positioning underscores their integral roles in cellular organization and function.

physics.bio-ph↗

arXiv2024

Entropy-driven decision-making dynamics sheds light on the emergence of the "paradox of choice"

Decision making is the cognitive process of selecting a course of action among multiple alternatives. As the decision maker belongs to a complex microenvironment (which contains multiple decision makers), has to make a decision where multiple options are present which often leads to a phenomenon known as the "paradox of choices". The latter refers to the case where too many options can lead to negative outcomes, such as increased uncertainty, decision paralysis, and frustration. Here, we employ an entropy driven mechanism within a statistical physics framework to explain the premises of the paradox. In turn, we focus on the emergence of a collective "paradox of choice", in the case of interacting decision-making agents, quantified as the decision synchronization time. Our findings reveal a trade-off between synchronization time and the sensing radius, indicating the optimal conditions for information transfer among group members, which significantly depends the individual sensitivity parameters. Interestingly, when agents sense their microenvironment in a biased way or their decisions are influenced by their past choices, then the collective "paradox of choice" does not occur. In a nutshell, our theory offers a low-dimensional and unified statistical explanation of the "paradox of choice" at the individual and at the collective level.

physics.soc-ph↗

arXiv2023

Cell decision-making through the lens of Bayesian learning

Cell decision-making refers to the process by which cells gather information from their local microenvironment and regulate their internal states to create appropriate responses. Microenvironmental cell sensing plays a key role in this process. Our hypothesis is that cell decision-making regulation is dictated by Bayesian learning. In this article, we explore the implications of this hypothesis for internal state temporal evolution. By using a timescale separation between internal and external variables on the mesoscopic scale, we derive a hierarchical Fokker-Planck equation for cell-microenvironment dynamics. By combining this with the Bayesian learning hypothesis, we find that changes in microenvironmental entropy dominate cell state probability distribution. Finally, we use these ideas to understand how cell sensing impacts cell decision-making. Notably, our formalism allows us to understand cell state dynamics even without exact biochemical information about cell sensing processes by considering a few key parameters.

physics.bio-ph↗

arXiv2020

Entropy-driven cell-decision making predicts fluid-to-solid transition in multicellular systems

Cellular decision making allows cells to assume functionally different phenotypes in response to microenvironmental cues, without genetic change. It is an open question, how individual cell decisions influence the dynamics at the tissue level. Here, we study spatio-temporal pattern formation in a population of cells exhibiting phenotypic plasticity, which is a paradigm of cell decision making. We focus on the migration/resting and the migration/proliferation plasticity which underly the epithelial-mesenchymal transition (EMT) and the go or grow dichotomy. We assume that cells change their phenotype in order to minimize their microenvironmental entropy (LEUP: Least microEnvironmental Uncertainty Principle) and study the impact of the LEUP-driven migration/resting and migration/proliferation plasticity on the corresponding multicellular spatio-temporal dynamics with a stochastic cell-based mathematical model for the spatio-temporal dynamics of the cell phenotypes. In the case of the go or rest plasticity, a corresponding mean-field approximation allows to identify a bistable switching mechanism between a diffusive (fluid) and an epithelial (solid) tissue phase which depends on the sensitivity of the phenotypes to the environment. For the go or grow plasticity, we show the possibility of Turing pattern formation for the "solid" tissue phase and its relation with the parameters of the LEUP-driven cell decisions.

physics.bio-ph↗

arXiv2016

A Mathematical Model of Cell Reprogramming due to Intermediate Differential Regulator's Regulations

In this paper I have given a mathematical model of Cell reprogramming from a different contexts. Here I considered there is a delay in differential regulator rate equations due to intermediate regulator's regulations. At first I gave some basic mathematical models by Ferell Jr.[2] of reprogramming and after that I gave mathematical model of cell reprogramming by Mithun Mitra[4]. In the last section I contributed a mathematical model of cell reprogramming from intermediate steps regulations and tried to find the critical point of pluripotent cell.

q-bio.CB↗