In this paper we compare three frameworks for modeling flows of complex fluids: (i) local conservations of mass, momentum and energy, (ii) GENERIC, and (iii) Onsager principle. The first is based on the mass, momentum, and energy conservation implied by mechanics, the second on the observed approach of externally unforced fluids to equilibrium states at which their behavior is well described by equilibrium thermodynamics, and the third on the minimal resistance to external influences. The comparison is illustrated on isothermal and incompressible polymeric fluids.

We study the resurgent structure of the topological string dual to 2d $U(N)$ Yang-Mills on torus. We find closed form formulas for instanton amplitudes up to arbitrarily high instanton orders, based on which we propose the non-perturbative partition function including contributions from all the real instantons, which is real for positive modulus and string coupling. We also explore complex instantons and find two infinite towers of them. We expect them to correspond to BPS states in type II string.

Industry adoption of chiplets has been growing as chiplets are a cost-effective option for making large, high-performance systems. Consequently, partitioning large systems into chiplets is increasingly important. In this work, we introduce ChipletPart, a cost-driven 2.5D system partitioner that addresses the constraints of chiplet systems, including complex objective functions, limited reach of inter-chiplet I/O transceivers, and the assignment of heterogeneous manufacturing technologies to different chiplets. ChipletPart integrates a sophisticated chiplet cost model with a genetic algorithm (GA)-based technology assignment and partitioning methodology, along with a simulated annealing (SA)-based chiplet floorplanner. Our results show that ChipletPart: (i) reduces chiplet cost by up to 58% (20% geometric mean) compared to state-of-the-art min-cut partitioners, which often yield floorplan-infeasible solutions; (ii) generates partitions with up to 47% (6% geometric mean) lower cost compared to the prior work Floorplet; (iii) reduces chiplet cost up to 48% (30% geometric mean) compared to Chipletizer, while consistently producing I/O-feasible chiplet solutions across all testcases; and (iv) for the testcases we study, heterogeneous integration reduces cost by up to 43% (15% geometric mean) compared to homogeneous implementations. Additionally, we explore Bayesian optimization (BO) for finding low cost and floorplan-feasible chiplet solutions with technology assignments. On some testcases, our BO framework achieves better system cost (up to 5.3% improvement) with higher runtime overhead (up to 4x) compared to our GA-based framework. We also present case studies that show how changes in packaging and inter-chiplet signaling technologies can affect partitioning solutions. Finally, ChipletPart, the underlying cost model, and our testcase generator are available as open-source tools.

In this work, we investigate the relativistic structure of white dwarfs (WDs) within the framework of modified gravity theory $f(R, T, L_m) = R + αT L_m$, which introduces a non-minimal coupling between matter and curvature. Using a realistic equation of state (EoS) that includes contributions from a relativistic degenerate electron gas and ionic lattice effects, we solve the modified Tolman-Oppenheimer-Volkoff (TOV) equations for two standard choices of the matter Lagrangian density: $L_m = p$ and $L_m = -ρ$. We show that the extra $αTL_m$ term significantly alters the mass-radius relation of WDs, especially at high central densities $( ρ_c \gtrsim 10^8 - 10^9\,\rm g/cm^3)$, allowing for stable super-Chandrasekhar configurations. In particular, depending on the sign and magnitude of the parameter $α$, the maximum mass can increase or decrease, and in some regimes, the usual critical point indicating the transition from stability to instability disappears. Our findings suggest that $f(R,T,L_m)$ gravity provides a viable framework to explain the existence of massive WDs beyond the classical Chandrasekhar limit. Using Bayesian inference with WD observational data, we further constrain the coupling parameter $α$ for the two choices of the Lagrangian density $L_m$.

In this study, we propose a robust control strategy for a counter-current heat exchanger. The primary objective is to regulate the outlet temperature of one fluid stream by manipulating the flow rate of the second counter-current fluid stream. By leveraging the energy balance equations, we develop a structured bilinear system model derived by using a uniform spatial discretization of each stream into a cascade of homogeneous volumes and by considering the heat transfer and convective phenomena within the exchanger. We introduce two control strategies: (i) an output feedback controller incorporating a state observer and (ii) a purely integral control law. The effectiveness of the proposed control strategy is validated through real experiments on a real heat exchanger.

We study a bulk-surface Cahn--Hilliard model with non-degenerate mobility and singular potentials in two dimensions. Following the ideas of the recent work by Conti, Galimberti, Gatti, and Giorgini [Calc. Var. Partial Differential Equations, 64(3):Paper No. 87, 32, 2025] for the Cahn--Hilliard equation with homogeneous Neumann boundary conditions, we show the uniqueness of weak solutions together with a continuous dependence estimate for sufficiently regular mobility functions. Next, under weaker assumptions on the mobility functions, we show the existence of a weak solution that exhibits the propagation of uniform-in-time regularity and satisfies the instantaneous separation property. Lastly, we consider the long-time behavior and prove that the unique weak solution converges to a solution of the stationary bulk-surface Cahn--Hilliard equation. Our approach for the uniqueness proof relies on a new well-posedness and regularity theory for a bulk-surface elliptic system with non-constant coefficients, which may be of independent interest.

The electric double layer (EDL) that forms at the interface between metals and ionic solutions is at the heart of various energy technologies. Recent experimental data have challenged our traditional understanding of the EDL charging behavior, which is based on mean-field Gouy-Chapman-Stern-type (GCS) models. In this article, we present a classical theory for the EDL, derived from first-principles statistical mechanics, that accounts for electron-ion correlation effects using the method of image charges and systematically extends beyond the mean-field level. Such electron-ion correlations introduce an additional interaction between the metal surface and electrolyte ions, significantly altering the EDL structure. Our theory, valid in the limit of dilute electrolyte solutions and weakly charged metal surfaces, achieves quantitative agreement with experimental capacitance data across a wide range of electrode materials and electrolyte solvents, and thus resolves long-standing questions on the origin of discrepancies to GCS predictions. Thereby, the framework conceptually unifies the processes of double-layer charging and ion adsorption (electrosorption), which are typically considered as distinct phenomena, but are shown to be manifestations of the same fundamental electrostatic principles.

We consider an operator-differential expression of the form $$ \ell y=\frac{d^m}{dx^m}\Big(By^{(n)}+Cy\Big), \quad 0<x<1, $$ where $B$ is a linear bounded invertible operator, while $C$ is some finite-dimensional linear operator relatively bounded to the operator of $n$-fold differentiation. To such a form, we can reduce, in particular, various singular differential expressions with the coefficients in negative Sobolev spaces, which creates an alternative to their regularization. In the case when $B$ is an integral Volterra operator of the second kind with a continuous kernel vanishing at the diagonal, we establish completeness of the root functions of an operator generated by the expression $\ell y$ and irregular semi-separated boundary conditions.

We consider a class of spontaneously broken $SU(2)$ gauge theories with adjoint scalar and look for exact magnetic monopole solutions in the Bogomol'nyi-Prasad-Sommerfield (BPS) limit. We find that some of the resulting solutions exhibit a new internal degree of freedom (a moduli space parameter) that controls the energy density profile of the monopole while keeping the total energy (mass) constant.

The integration of active light-emitting elements into planar photonic circuits on a silicon nitride platform remains challenging due to material incompatibilities and high-temperature processing. Proposed hybrid method embeds monodisperse luminescent particles into lithographically defined wells above a 200 nm-thick silicon nitride taper coupler. A fabrication process involving wells etching, particle deposition, and planarization enables precise integration while maintaining waveguide integrity. When pumped at 950 nm with a diode laser, the device emits broadband radiation in the 1500-1600 nm range, covering the optical telecommunication C-band. Numerical simulations yield an average coupling efficiency of 0.25% into the fundamental waveguide mode, suggesting significant potential for further device optimization. The approach provides a scalable route for integrating broadband telecommunications emitters on a silicon nitride platform.

While twisted magnonic crystals (MCs) have recently gained attention for their intriguing linear phenomena, such as magnon flat bands, their nonlinear dynamics -- particularly the generation of magnonic frequency combs (MFCs) -- have remained largely unexplored. In this work, we demonstrate the creation of MFCs in twisted MCs using two-tone microwave excitation. We find that finite twist angles significantly enhance three-magnon interactions, driven by the non-collinear ground-state magnetic configuration induced by interlayer dipole-dipole interactions. The number of comb teeth exhibits a plateau-like dependence on the twist angle, with the plateau's width and height saturating as the excitation frequency of the propagating magnon mode increases. This behavior reveals an optimal range of twist angles and frequencies for achieving high-quality MFCs with a large number of comb teeth. Our findings deepen the understanding of nonlinear interactions in twisted MCs and highlight their potential for advancing moiré-based materials in information processing and high-precision metrology.

The problem of bistatic integrated sensing and communications over memoryless relay channels is considered, where destination concurrently decodes the message sent by the source and estimates unknown parameters from received signals with the help of a relay. A state-dependent discrete memoryless relay channel is considered to model this setup, and the fundamental limits of the communication-sensing performance tradeoff are characterized by the capacity-distortion function. An upper bound on the capacity-distortion function is derived, extending the cut-set bound results to address the sensing operation at the destination. A hybrid-partial-decode-and-compress-forward coding scheme is also proposed to facilitate source-relay cooperation for both message transmission and sensing, establishing a lower bound on the capacity-distortion function. It is found that the hybrid-partial-decode-and-compress-forward scheme achieves optimal sensing performance when the communication task is ignored. Furthermore, the upper and lower bounds are shown to coincide for three specific classes of relay channels. Numerical examples are provided to illustrate the communication-sensing tradeoff and demonstrate the benefits of integrated design.

Achieving high-fidelity qubit readout and reset while preserving qubit coherence is essential for quantum error correction and other advanced quantum algorithms. Here, we design and experimentally demonstrate a scalable architecture employing frequency-tunable nonlinear Purcell filters, enabling flexible readout and fast unconditional reset of multiple superconducting qubits. Our readout protocol dynamically adjusts the effective linewidth of the readout resonator through a tunable Purcell filter, optimizing the signal-to-noise ratio during measurement while suppressing photon noise during idle periods. We achieve a readout fidelity of $99.3\%$ without any quantum-limited amplifier, even with a small dispersive shift. Moreover, by leveraging a reset channel formed via the adjacent coupling between the filter and the coupler, we realize unconditional qubit reset of both leakage-induced $|2\rangle$ and $|1\rangle$ states within 200 ns and reset of the $|1\rangle$ state alone within 75 ns, with error rates $\leq 1\%$. The filter also mitigates both photon-induced dephasing and the Purcell effect, thereby preserving qubit coherence. This scalable Purcell filter architecture shows exceptional performance in qubit readout, reset, and protection, marking it as a promising hardware component for advancing fault-tolerant quantum computing systems.

Extreme mass-ratio inspirals (EMRIs), in which a solar mass compact object is whirling around a supermassive black hole, act as precise tracers of the spacetime geometry and astrophysical environment around the supermassive black hole. These systems are highly sensitive to even the smallest deviations from the vacuum general relativity scenario. However, detecting these signals requires highly accurate waveform modeling up to the first post-adiabatic order, incorporating self-force effects, system parameters, and environmental influences. In this paper, we focus on the impact of dark matter on gravitational waveforms. Cold dark matter in galactic centers can be redistributed by the gravitational pull of a supermassive black hole, forming a dense, spike-like profile. When an EMRI evolves in such an environment, the interaction between the binary and the surrounding dark matter can leave distinctive imprints on the emitted waveform, and thus offer a novel way to probe the nature and distribution of dark matter. We specifically examine how dark matter modifies the background spacetime. By treating these modifications perturbatively, we present a framework to incorporate dark matter environmental effects into gravitational waveform modeling at the first post-adiabatic order.

Picture an experimental scenario where a closed quantum system, evolving through a time-independent Hamiltonian, is subject to a demolition measurement at a chosen time. The Hamiltonian, the measured observables, the initial state of the physical system and even its Hilbert space dimension are unknown; we nonetheless assume a promise or constraint on the energy distribution of the state. In this context we find that, for many natural energy constraints, the set of feasible time series or datasets can be characterized efficiently. Furthermore, under the assumption of a bounded energy spectrum, we prove that there exist ``self-testing'' datasets, whose approximate realization singles out specific Hamiltonians, states and measurement operators. Investigating to what extent the extrapolation of past measurement data is possible in this framework, we identify energy-constrained physical systems for which a non-trivial prediction at time $τ$ requires a precision in the measurement data superexponential in $τ$. We also discover two extrapolation phenomena: ``aha! datasets", which drastically increase the predictability of the future statistics of an unrelated measurement; and ``fog banks": fairly simple datasets that exhibit complete unpredictability at some future time $τ$, but full predictability at a later time $τ'>τ$. Besides their relevance for quantum foundations, our results have applications in semi-device independent quantum communication, the simulation of complex quantum systems, and the design of optimal atomic clocks.

The Jiangmen Underground Neutrino Observatory (JUNO) employs a 20 kt liquid scintillator (LS) detector located 700 m underground. To meet its physics objectives, the LS must achieve an ultra-low $^{238}$U/$^{232}$Th content of 10$^{-17}$ g/g. Given that airborne dust exhibits radioactivity about 12 orders of magnitude higher, exceptional cleanliness is essential during on-site installation. The total permissible dust mass in the 20 kt LS is only about 8 mg. To attain this, the acrylic vessel interior must comply with class 1,000 cleanliness. Pre-filling water spray cleaning improves cleanliness by roughly two orders of magnitude, requiring the overall environment to be maintained between class 10,000 and 100,000. At JUNO, a cleanroom management system has been implemented across the 120,000 m$^3$ underground experimental hall. Since May 2022, continuous laser particle monitoring has consistently achieved an average cleanliness class of 74,000. Furthermore, we developed a method to directly measure $^{238}$U/$^{232}$Th deposition rates on detector surfaces. Using ICP-MS, sensitivity reaches sub-ppt levels ($<$10$^{-12}$ g/g), enabling effective cleanliness control and assessment of external contamination during detector construction.

Competing superconducting (SC) and density-wave orders are of key importance in generating unconventional superconductivity and emergent electronic responses. Quasi-one-dimensional models provide insight into these competing orders and suggest higher-dimensional realizations through coupled-wire constructions, but analysis of such systems remains limited. Recent studies suggest that double helical edge states (DHESs) in double quantum spin Hall insulators (DQSHIs) form a two-channel Luttinger liquid that exhibits SC and spin density wave (SDW) phases and their $π$-junction analogs. Here, we analyze weakly coupled DHESs from the surface of a periodically stacked layered structure consisting of DQSHIs and dielectrics, where inter-edge interactions approximately develop a tunable helical sliding Luttinger liquid (HSLL) order. Using a renormalization-group analysis, we construct phase diagrams and identify a regime of HSLL parameters that favor competing two-dimensional $π$-SC and $π$-SDW orders. We identify parameter regimes where the competing orders could be realized experimentally in nanoscale devices. Our study suggests a promising materials platform for exploring tunable $π$-SC and $π$-SDW orders in double quantum spin Hall superlattices.

This paper introduces a new method for solving the planar heat equation based on the Lightning Method. The lightning method is a recent development in the numerical solution of linear PDEs which expresses solutions using sums of polynomials and rational functions, or more generally as sums of fundamental solutions. The method is particularly well suited to handle domains with sharp corners where solution singularities are present. Boundary conditions are formed on a set of collocation points which is then solved as an overdetermined linear system. The approach of the present work is to utilize the Laplace transform to obtain a modified Helmholtz equation which is solved by an application of the lightning method. The numerical inversion of the Laplace transform is then performed by means of Talbot integration. Our validation of the method against existing results and multiple challenging test problems shows the method attains spectral accuracy with root-exponential convergence while being robust across a wide range of time intervals and adaptable to a variety of geometric scenarios.

Three-dimensional topological insulators possess topologically protected surface states with spin-momentum locking, which enable spin-charge-current interconversion (SCI) by the inverse Edelstein effect (IEE). However, it remains experimentally challenging to separate the surface-related IEE from the bulk-type inverse spin Hall effect (ISHE). Here, we search for distinct time-domain signatures of the two SCI phenomena in a $\mathcal{F}$|TI model stack of a ferromagnetic-metal layer $\mathcal{F}$ (Co and Fe) and a topological-insulator layer TI (Bi$_2$Te$_3$, SnBi$_2$Te$_4$ and Bi$_{1-x}$Sb$_x$ with $x$ = 0.15 and 0.3), where the focus is on Bi$_2$Te$_3$. A femtosecond laser pulse serves to induce a transient spin voltage $μ_s^{\mathcal{F}}$ in $\mathcal{F}$ and, thus, drive an ultrafast spin current out of $\mathcal{F}$. SCI results in a transverse charge current with a sheet density $I_c$ that is detected by sampling the emitted terahertz electric field. Analysis of the dynamics of $I_c(t)$ vs time $t$ relative to $μ_s^{\mathcal{F}}(t)$ reveals two components with distinct time scales: (i) a quasi-instantaneous response and (ii) a longer-lived response with a relaxation time of 270 fs, which is independent of the chosen $\mathcal{F}$ material. Component (i) is consistently ascribed to the ISHE. In contrast, we interpret component (ii) as a signature of interfacial spin accumulation and the IEE at the $\mathcal{F}$/Bi$_2$Te$_3$ interface, with a fraction of $< 10^{-2}$ of the incident spins participating. This assignment is fully consistent with respect to its dynamics and magnitude. We rate other possible signal contributions, such as spin trapping in intermediate states, as less likely. Our results show that the femtosecond dynamics of photocurrents provide important insights into the mechanisms of spin transport and SCI in $\mathcal{F}$|TI stacks.

We sharpen the moment comparison inequalities with sharp constants for sums of random vectors uniform on Euclidean spheres, providing a deficit term (optimal in high dimensions).

In this paper, we develop algorithms for computing the recurrence coefficients corresponding to multiple orthogonal polynomials on the step-line. We reformulate the problem as an inverse eigenvalue problem, which can be solved using numerical linear algebra techniques. We consider two approaches: the first is based on the link with block Krylov subspaces and results in a biorthogonal Lanczos process with multiple starting vectors; the second consists of applying a sequence of Gaussian eliminations on a diagonal matrix to construct the banded Hessenberg matrix containing the recurrence coefficients. We analyze the accuracy and stability of the algorithms with numerical experiments on the ill-conditioned inverse eigenvalue problemshave related to Kravchuk and Hahn polynomials, as well as on other better conditioned examples.

Coherent magnon modes in a truly three-dimensional (3D) magnonic crystal have not yet been investigated. This scientific gap exists despite the numerous theoretical predictions about miniband formation and edge modes with topological protection. Such properties are key to advance nanomagnonics for ultrafast data processing. In this work, we use a scalable nanotechnology and integrate a 3D magnonic crystal to an on-chip microresonator. It was fabricated by two-photon lithography of a 3D woodpile structure and atomic layer deposition of 30-nm-thick nickel. Operated near 14 and 24~GHz, the microresonator output revealed numerous coherent magnons with distinct angular dependencies reflecting the underlying face-centred cubic lattice. Micromagnetic simulations show that the edge modes are localised in curved nanocaps and robust against changes in field orientation. Along an edge, they exhibit an unexpected phase evolution. Our findings advance functional microwave circuits with 3D magnonic crystals and fuel their visionary prospects of edge-dominated magnon modes.

Biological systems operate under persistent noise, which can alter system states and induce transitions between attractors. Here, we study the attractor dynamics of Boolean networks focusing on the transitions between attractors induced by noise. By computing transition probabilities between attractors, we present methods at the attractor level to determine dominance, stability, and diversity of attractors, and systematically compare local and global noise. Whereas global noise leads to attractor behavior dictated primarily by basin sizes, local noise produces structured transition patterns characterized by enhanced stability, non-trivial dominance patterns, and broader exploration of the attractor space. Our work offers insight into the dynamics of attractors, showing the importance of transition patterns under noise.

In this paper we consider the numerical approximation of a semilinear reaction-diffusion model problem (PDEs) by means of reduced order methods (ROMs) based on proper orthogonal decomposition (POD). We focus on the time integration of the fully discrete reduced order model. Most of the analysis in the literature has been carried out for the implicit Euler method as time integrator. We integrate in time the reduced order model with the BDF-q time stepping ($1\le q\le 5$) and prove optimal rate of convergence of order $q$ in time. Our set of snapshots is obtained from finite element approximations to the original model problem computed at different times. These finite element approximations can be obtained with any time integrator. The POD method is based on first order difference quotients of the snapshots. The reason for doing this is twofold. On the one hand, the use of difference quotients allow us to provide pointwise-in-time error bounds. On the other, the use of difference quotients is essential to get the expected rate $q$ in time since we apply that the BDF-q time stepping, $1\le q\le 5$, can be written as a linear combination of first order difference quotients.

We derive an integral expression $G(z)$ for the reciprocal gamma function, $1/Γ(z)=G(z)/π$, that is valid for all $z\in\mathbb{C}$, without the need for analytic continuation. The same integral avoids the singularities of the gamma function and satisfies $G(1-z)=Γ(z)\sin(πz)$ for all $z\in\mathbb{C}$.

This paper introduces two variational inference approaches for infinite-dimensional inverse problems, developed through gradient descent with a constant learning rate. The proposed methods enable efficient approximate sampling from the target posterior distribution using a constant-rate stochastic gradient descent (cSGD) iteration. Specifically, we introduce a randomization strategy that incorporates stochastic gradient noise, allowing the cSGD iteration to be viewed as a discrete-time process. This transformation establishes key relationships between the covariance operators of the approximate and true posterior distributions, thereby validating cSGD as a variational inference method. We also investigate the regularization properties of the cSGD iteration and provide a theoretical analysis of the discretization error between the approximated posterior mean and the true background function. Building on this framework, we develop a preconditioned version of cSGD to further improve sampling efficiency. Finally, we apply the proposed methods to two practical inverse problems: one governed by a simple smooth equation and the other by the steady-state Darcy flow equation. Numerical results confirm our theoretical findings and compare the sampling performance of the two approaches for solving linear and non-linear inverse problems.

We report transport studies of hybrid Josephson junctions based on semiconducting InAs nanowires with fully overlapping epitaxial ferromagnetic insulator EuS and superconducting Al partial shells. Current-biased measurements reveal a hysteretic superconducting window with a sizable supercurrent near the coercive field of the ferromagnetic insulator, accompanied by multiple Andreev reflections. Tunneling spectroscopy shows a superconducting gap characterized by three peaks, which we attribute to tunneling between exchange-split superconductors. A theoretical model reproduces the observed features and indicates that spin mixing, driven by sizable spin-orbit coupling, is essential to their formation. Our results demonstrate proximity-induced superconductivity through a ferromagnetic insulator and establish a new platform for exploring spin-triplet pairing.

The complex geometry of the Ly$α$ forest data has motivated the use of various two-point statistics as alternatives to the three-dimensional power spectrum ($P_{\mathrm{3D}}$), which carries cosmological information in Fourier space. On large scales, the three-dimensional correlation function ($ξ_\mathrm{3D}$) has provided robust measurements of the baryon acoustic oscillation (BAO) scale at 150~Mpc. On smaller scales, the one-dimensional power spectrum, $P_{\mathrm{1D}}(k_\|)$, has been the primary tool for extracting information. At the same time, the cross-spectrum, $P_\times(θ, k_\|)$, has been introduced to incorporate angular information without the complications caused by survey window functions. We propose an analytical forward-modeling framework to reconstruct $P_{\mathrm{3D}}$ from all these observables, based on the mathematical relation between them and $P_{\mathrm{3D}}$. We demonstrate the performance of our method using a hypothetical mock data vector representative of future Dark Energy Spectroscopic Instrument (DESI) measurements. We show that the monopole of $P_{\mathrm{3D}}$ can be reconstructed in 25 $k$ bins between $0.07~\mathrm{Mpc}^{-1}$ and $1.8~\mathrm{Mpc}^{-1}$, achieving an average precision of $σ_P/P=13\%$ across the bins. Our method can serve as an intermediary for consistency checks, though it is not intended to replace direct $P_{\mathrm{3D}}$ estimation.

Classical computers can simulate models of quantum computation with restricted input states. The identification of such states can sharpen the boundary between quantum and classical computations. Previous works describe simulable states of odd-dimensional systems. Here, we further our understanding of systems of qubits. We do so by casting a real-quantum-bit model of computation in terms of a Kirkwood-Dirac (KD) quasiprobability distribution. Algorithms, throughout which this distribution is a proper (positive) probability distribution can be simulated efficiently on a classical computer. We leverage recent results on the geometry of the set of KD-positive states to construct previously unknown classically-simulable (bound) states. Finally, we show that KD nonpositivity is a resource monotone for quantum computation, establishing KD nonpositivity as a necessary resource for computational quantum advantage.

In this work, we aim to reconcile several apparently contradictory observations in market microstructure: is the famous "square-root law" of metaorder impact, which decays with time, compatible with the random-walk nature of prices and the linear impact of order imbalances? Can one entirely explain the volatility of prices as resulting from the flow of uninformed metaorders that mechanically impact them? We introduce a new theoretical framework to describe metaorders with different signs, sizes and durations, which all impact prices as a square-root of volume but with a subsequent time decay. We show that, as in the original propagator model, price diffusion is ensured by the long memory of cross-correlations between metaorders. In order to account for the effect of strongly fluctuating volumes q of individual trades, we further introduce two q-dependent exponents, which allow us to describe how the moments of generalized volume imbalance and the correlation between price changes and generalized order flow imbalance scale with T. We predict in particular that the corresponding power-laws depend in a non-monotonic fashion on a parameter a, which allows one to put the same weight on all child orders or to overweight large ones, a behaviour that is clearly borne out by empirical data. We also predict that the correlation between price changes and volume imbalances should display a maximum as a function of a, which again matches observations. Such noteworthy agreement between theory and data suggests that our framework correctly captures the basic mechanism at the heart of price formation, namely the average impact of metaorders. We argue that our results support the "Order-Driven" theory of excess volatility, and are at odds with the idea that a "Fundamental" component accounts for a large share of the volatility of financial markets.