Showing 1–12 of 12 results for author: Jazayeri, M

One-way Optomechanical Interaction between Nanoparticles

Authors: Amir M. Jazayeri, Sohila Abdelhafiz, Aristide Dogariu

Abstract: Within a closed system, physical interactions are reciprocal. However, the effective interaction between two entities of an open system may not obey reciprocity. Here, we describe a non-reciprocal interaction between nanoparticles which is one-way, almost insensitive to the interparticle distance, and scalable to many particles. The interaction we propose is based on the non-conservative optical f… ▽ More Within a closed system, physical interactions are reciprocal. However, the effective interaction between two entities of an open system may not obey reciprocity. Here, we describe a non-reciprocal interaction between nanoparticles which is one-way, almost insensitive to the interparticle distance, and scalable to many particles. The interaction we propose is based on the non-conservative optical forces between two nanoparticles with highly directional scattering patterns. However, we elucidate that scattering patterns can in general be very misleading about the interparticle forces. We introduce zeroth- and first-order non-reciprocity factors to precisely quantify the merits of any optomechanical interaction between nanoparticles. Our proposed one-way interaction could constitute an important step in the realization of mesoscopic heat pumps and refrigerators, the study of non-equilibrium systems, and the simulation of non-Hermitian quantum models. △ Less

Submitted 31 January, 2024; originally announced February 2024.

Neural Foundations of Mental Simulation: Future Prediction of Latent Representations on Dynamic Scenes

Authors: Aran Nayebi, Rishi Rajalingham, Mehrdad Jazayeri, Guangyu Robert Yang

Abstract: Humans and animals have a rich and flexible understanding of the physical world, which enables them to infer the underlying dynamical trajectories of objects and events, plausible future states, and use that to plan and anticipate the consequences of actions. However, the neural mechanisms underlying these computations are unclear. We combine a goal-driven modeling approach with dense neurophysiol… ▽ More Humans and animals have a rich and flexible understanding of the physical world, which enables them to infer the underlying dynamical trajectories of objects and events, plausible future states, and use that to plan and anticipate the consequences of actions. However, the neural mechanisms underlying these computations are unclear. We combine a goal-driven modeling approach with dense neurophysiological data and high-throughput human behavioral readouts to directly im**e on this question. Specifically, we construct and evaluate several classes of sensory-cognitive networks to predict the future state of rich, ethologically-relevant environments, ranging from self-supervised end-to-end models with pixel-wise or object-centric objectives, to models that future predict in the latent space of purely static image-based or dynamic video-based pretrained foundation models. We find strong differentiation across these model classes in their ability to predict neural and behavioral data both within and across diverse environments. In particular, we find that neural responses are currently best predicted by models trained to predict the future state of their environment in the latent space of pretrained foundation models optimized for dynamic scenes in a self-supervised manner. Notably, models that future predict in the latent space of video foundation models that are optimized to support a diverse range of sensorimotor tasks, reasonably match both human behavioral error patterns and neural dynamics across all environmental scenarios that we were able to test. Overall, these findings suggest that the neural mechanisms and behaviors of primate mental simulation are thus far most consistent with being optimized to future predict on dynamic, reusable visual representations that are useful for Embodied AI more generally. △ Less

Submitted 25 October, 2023; v1 submitted 19 May, 2023; originally announced May 2023.

Comments: 20 pages, 10 figures, NeurIPS 2023 Camera Ready Version (spotlight)

Modeling Human Eye Movements with Neural Networks in a Maze-Solving Task

Authors: Jason Li, Nicholas Watters, Yingting, Wang, Hansem Sohn, Mehrdad Jazayeri

Abstract: From smoothly pursuing moving objects to rapidly shifting gazes during visual search, humans employ a wide variety of eye movement strategies in different contexts. While eye movements provide a rich window into mental processes, building generative models of eye movements is notoriously difficult, and to date the computational objectives guiding eye movements remain largely a mystery. In this wor… ▽ More From smoothly pursuing moving objects to rapidly shifting gazes during visual search, humans employ a wide variety of eye movement strategies in different contexts. While eye movements provide a rich window into mental processes, building generative models of eye movements is notoriously difficult, and to date the computational objectives guiding eye movements remain largely a mystery. In this work, we tackled these problems in the context of a canonical spatial planning task, maze-solving. We collected eye movement data from human subjects and built deep generative models of eye movements using a novel differentiable architecture for gaze fixations and gaze shifts. We found that human eye movements are best predicted by a model that is optimized not to perform the task as efficiently as possible but instead to run an internal simulation of an object traversing the maze. This not only provides a generative model of eye movements in this task but also suggests a computational theory for how humans solve the task, namely that humans use mental simulation. △ Less

Submitted 20 December, 2022; originally announced December 2022.

Fixed points on band structures of non-Hermitian models: Extended states in the bandgap and ideal superluminal tunneling

Authors: Amir M. Jazayeri

Abstract: Space-time reflection symmetry (PT symmetry) in non-Hermitian electronic models has drawn much attention over the past decade mainly because it guarantees that the band structures calculated under open boundary conditions be the same as those calculated under periodic boundary conditions. PT symmetry in electromagnetic (EM) models, which are usually borrowed from electronic models, has also been o… ▽ More Space-time reflection symmetry (PT symmetry) in non-Hermitian electronic models has drawn much attention over the past decade mainly because it guarantees that the band structures calculated under open boundary conditions be the same as those calculated under periodic boundary conditions. PT symmetry in electromagnetic (EM) models, which are usually borrowed from electronic models, has also been of immense interest mainly because it leads to 'exceptional' parameter values below which non-Hermitian operators have real eigenvalues, although PT symmetry is not the sole symmetry which allows such exceptional parameter values. In this article, we examine one-dimensional PT-symmetric non-Hermitian EM models to introduce novel concepts and phenomena. We introduce the band-structure concept of 'fixed points', which leads to 'bidirectional' reflection zeros in the corresponding finite structures, contrary to a common belief about the EM structures with PT symmetry (and without P and T symmetries). Some of the fixed points manifest themselves as what we name 'extended states in the bandgap' on the band structure while some other fixed points are the 'turning points' of the band structure. The extended states in the band gap are in fact the dual of the well-known 'bound states in the continuum' while the turning points allow us to observe 'ideal' superluminal tunneling in the corresponding finite structures. By 'ideal' superluminal tunneling we mean the case where not only the transmission coefficient has an almost uniform phase over a broad bandwidth but also the magnitudes of the transmission and reflection coefficients are almost equal to unity and zero, respectively, over the bandwidth. △ Less

Submitted 16 April, 2023; v1 submitted 10 August, 2022; originally announced August 2022.

Journal ref: Physical Review B 107.14 (2023) 144302

Neural Latents Benchmark '21: Evaluating latent variable models of neural population activity

Authors: Felix Pei, Joel Ye, David Zoltowski, Anqi Wu, Raeed H. Chowdhury, Hansem Sohn, Joseph E. O'Doherty, Krishna V. Shenoy, Matthew T. Kaufman, Mark Churchland, Mehrdad Jazayeri, Lee E. Miller, Jonathan Pillow, Il Memming Park, Eva L. Dyer, Chethan Pandarinath

Abstract: Advances in neural recording present increasing opportunities to study neural activity in unprecedented detail. Latent variable models (LVMs) are promising tools for analyzing this rich activity across diverse neural systems and behaviors, as LVMs do not depend on known relationships between the activity and external experimental variables. However, progress with LVMs for neuronal population activ… ▽ More Advances in neural recording present increasing opportunities to study neural activity in unprecedented detail. Latent variable models (LVMs) are promising tools for analyzing this rich activity across diverse neural systems and behaviors, as LVMs do not depend on known relationships between the activity and external experimental variables. However, progress with LVMs for neuronal population activity is currently impeded by a lack of standardization, resulting in methods being developed and compared in an ad hoc manner. To coordinate these modeling efforts, we introduce a benchmark suite for latent variable modeling of neural population activity. We curate four datasets of neural spiking activity from cognitive, sensory, and motor areas to promote models that apply to the wide variety of activity seen across these areas. We identify unsupervised evaluation as a common framework for evaluating models across datasets, and apply several baselines that demonstrate benchmark diversity. We release this benchmark through EvalAI. http://neurallatents.github.io △ Less

Submitted 17 January, 2022; v1 submitted 9 September, 2021; originally announced September 2021.

Interpreting neural computations by examining intrinsic and embedding dimensionality of neural activity

Authors: Mehrdad Jazayeri, Srdjan Ostojic

Abstract: The ongoing exponential rise in recording capacity calls for new approaches for analysing and interpreting neural data. Effective dimensionality has emerged as an important property of neural activity across populations of neurons, yet different studies rely on different definitions and interpretations of this quantity. Here we focus on intrinsic and embedding dimensionality, and discuss how they… ▽ More The ongoing exponential rise in recording capacity calls for new approaches for analysing and interpreting neural data. Effective dimensionality has emerged as an important property of neural activity across populations of neurons, yet different studies rely on different definitions and interpretations of this quantity. Here we focus on intrinsic and embedding dimensionality, and discuss how they might reveal computational principles from data. Reviewing recent works, we propose that the intrinsic dimensionality reflects information about the latent variables encoded in collective activity, while embedding dimensionality reveals the manner in which this information is processed. We conclude by highlighting the role of network models as an ideal substrate for testing more specifically various hypotheses on the computational principles reflected through intrinsic and embedding dimensionality. △ Less

Submitted 27 August, 2021; v1 submitted 8 July, 2021; originally announced July 2021.

Modular Object-Oriented Games: A Task Framework for Reinforcement Learning, Psychology, and Neuroscience

Authors: Nicholas Watters, Joshua Tenenbaum, Mehrdad Jazayeri

Abstract: In recent years, trends towards studying simulated games have gained momentum in the fields of artificial intelligence, cognitive science, psychology, and neuroscience. The intersections of these fields have also grown recently, as researchers increasing study such games using both artificial agents and human or animal subjects. However, implementing games can be a time-consuming endeavor and may… ▽ More In recent years, trends towards studying simulated games have gained momentum in the fields of artificial intelligence, cognitive science, psychology, and neuroscience. The intersections of these fields have also grown recently, as researchers increasing study such games using both artificial agents and human or animal subjects. However, implementing games can be a time-consuming endeavor and may require a researcher to grapple with complex codebases that are not easily customized. Furthermore, interdisciplinary researchers studying some combination of artificial intelligence, human psychology, and animal neurophysiology face additional challenges, because existing platforms are designed for only one of these domains. Here we introduce Modular Object-Oriented Games, a Python task framework that is lightweight, flexible, customizable, and designed for use by machine learning, psychology, and neurophysiology researchers. △ Less

Submitted 24 February, 2021; originally announced February 2021.

Critical Ambient Pressure and Critical Cooling Rate in Optomechanics of Electromagnetically Levitated Nanoparticles

Authors: Amir M. Jazayeri

Abstract: The concept of critical ambient pressure is introduced. The particle escapes from its trap when the ambient pressure becomes comparable with or smaller than a critical value, even if the particle motion is cooled by one of the feedback (or cavity) cooling schemes realized so far. The critical ambient pressure may be so small that it is not a limiting factor in ground-state cooling, but critical fe… ▽ More The concept of critical ambient pressure is introduced. The particle escapes from its trap when the ambient pressure becomes comparable with or smaller than a critical value, even if the particle motion is cooled by one of the feedback (or cavity) cooling schemes realized so far. The critical ambient pressure may be so small that it is not a limiting factor in ground-state cooling, but critical feedback cooling rates, which are also introduced, are limiting factors. The particle escapes from its trap if any of the feedback cooling rates becomes comparable with or larger than its critical value. Critical feedback cooling rate is different from the well-known manifestation of the measurement noise. The critical feedback cooling rate corresponding to a certain component of the particle motion is usually smaller than the optimum feedback cooling rate at which the standard quantum limit happens unless that component is cooled by the Coulomb force (instead of the optical gradient force). Also, given that the measurement noise for the z component is smaller than the measurement noises for the other two components (assuming that the beam illuminating the particle for photodetection propagates parallel to the z axis), the feedback scheme in which the z component of the particle motion is cooled by the Coulomb force has the best performance. This conclusion is in agreement with the experimental results published after writing the first version of this paper. The dependence of the critical ambient pressure, the critical feedback cooling rates, and the minimum achievable mean phonon numbers on the parameters of the system is derived, and can be verified experimentally. The insights into and the subtle points about the EM force, the EM force fluctuations, and the measurement noise that are presented might be useful in many systems besides those examined here. △ Less

Submitted 27 October, 2021; v1 submitted 7 June, 2019; originally announced June 2019.

Comment on "Inclusion of the backaction term in the total optical force exerted upon Rayleigh particles in nonresonant structures"

Authors: Amir M. Jazayeri

Abstract: The conventional dipole approximation (CDA) assumes that the EM fields a small particle generates in the presence of its surrounding material bodies are equal to the EM fields a point-like dipole generates in the absence of the material bodies. The authors of [Phys. Rev. A 98, 013806 (2018)] investigate a modified dipole approximation (MDA), which assumes that the EM fields the particle generates… ▽ More The conventional dipole approximation (CDA) assumes that the EM fields a small particle generates in the presence of its surrounding material bodies are equal to the EM fields a point-like dipole generates in the absence of the material bodies. The authors of [Phys. Rev. A 98, 013806 (2018)] investigate a modified dipole approximation (MDA), which assumes that the EM fields the particle generates are equal to the EM fields a point-like dipole generates in the presence of the material bodies. The authors interpret the approximate EM force under the MDA as the sum of four terms named 'generalized gradient force', 'generalized radiation pressure', 'generalized spin curl force', and 'new force term'. I show that such an interpretation is wrong and misleading: the generalized gradient force is not a gradient force; the generalized radiation pressure is not a radiation pressure; and the generalized spin curl force is not a spin curl force. In the case where the particle interacts with only one resonant EM mode of small enough linewidth, which is usually called 'self-induced back-action trap**' in the literature, I rectify some of the statements in their paper, and clear up some common misconceptions in the literature. In the case where the particle does not interact with any EM modes of small enough linewidth, I show that the numerical examples in their paper are inconclusive, and more importantly, the MDA is in principle inaccurate when the CDA is inaccurate. It should be noted that finding the approximate EM force under the MDA is computationally as hard as finding the exact EM force. △ Less

Submitted 15 October, 2019; v1 submitted 27 May, 2019; originally announced May 2019.

All-dielectric Structure for Trap** Nanoparticles via Light Funneling and Nanofocusing

Authors: Amir M. Jazayeri, Khashayar Mehrany

Abstract: We propose a dielectric structure which focuses the laser light well beyond the diffraction limit and thus considerably enhances the exerted optical trap** force upon dielectric nanoparticles. Although the structure supports a Fabry-Perot resonance, it actually acts as a nanoantenna in that the role of the resonance is to funnel the laser light into the structure. In comparison with the lens ill… ▽ More We propose a dielectric structure which focuses the laser light well beyond the diffraction limit and thus considerably enhances the exerted optical trap** force upon dielectric nanoparticles. Although the structure supports a Fabry-Perot resonance, it actually acts as a nanoantenna in that the role of the resonance is to funnel the laser light into the structure. In comparison with the lens illuminating the structure, the proposed structure offers roughly a 10000-fold enhancement in the trap** force - part of this enhancement comes from an 80-fold enhancement in the field intensity, and the remaining comes from a 130-fold enhancement in the normalized gradient of the field intensity (viz. the gradient of the field intensity divided by the field intensity). Also, the proposed structure offers roughly a 100-fold enhancement in the depth of the trap** potential. It is noteworthy that 'self-induced back-action trap**' (SIBA), which has recently been the focus of interest in the context of optical resonators, does not take place in the proposed resonator. In this paper, we also point out some misconceptions about SIBA together with some hitherto unappreciated subtleties of the dipole approximation. △ Less

Submitted 6 May, 2019; v1 submitted 30 January, 2017; originally announced January 2017.

Achievability of Quantum Regime in Optomechanics of Electromagnetically Levitated Nanoparticles: Inclusion of Gradient Force Fluctuations

Authors: Amir M. Jazayeri, Khashayar Mehrany

Abstract: Counterintuitively, experiments show that an electromagnetically levitated particle escapes from its trap when the ambient pressure is reduced below a certain level even if the particle's motion is cooled by a resonator-based or feedback-based mechanism. Here, we theoretically show that the ambient pressure must be kept well above a critical value arising from gradient force fluctuations (viz., fl… ▽ More Counterintuitively, experiments show that an electromagnetically levitated particle escapes from its trap when the ambient pressure is reduced below a certain level even if the particle's motion is cooled by a resonator-based or feedback-based mechanism. Here, we theoretically show that the ambient pressure must be kept well above a critical value arising from gradient force fluctuations (viz., fluctuations in part of the EM force whose Hamiltonian is quadratic in the position of the particle). Also, we consider other force fluctuations, and determine whether different realizations of feedback cooling are able to reach the ground state. In some realizations, the cooling rate must be kept well below a critical value arising from feedback-induced gradient force fluctuations. △ Less

Submitted 6 May, 2019; v1 submitted 3 February, 2016; originally announced February 2016.

Critical study and discrimination of different formulations of electromagnetic force density and consequent stress tensors inside matter

Authors: Amir M. Jazayeri, Khashayar Mehrany

Abstract: By examination of the exerted electromagnetic (EM) force on boundary of an object in a few examples, we look into the compatibility of the stress tensors corresponding to different formulae of the EM force density with special relativity. Ampere-Lorentz's formula of the EM force density is physically justifiable in that the electric field and the magnetic flux density act on the densities of the t… ▽ More By examination of the exerted electromagnetic (EM) force on boundary of an object in a few examples, we look into the compatibility of the stress tensors corresponding to different formulae of the EM force density with special relativity. Ampere-Lorentz's formula of the EM force density is physically justifiable in that the electric field and the magnetic flux density act on the densities of the total charges and the total currents, unlike Minkowski's formula which completely excludes the densities of the bounded charges and the bounded currents inside homogeneous media. Abraham's formula is fanciful and devoid of physical meaning. Einstein-Laub's formula seems to include the densities of the total charges and the total currents at first sight, but grou** the bounded charges and the bounded currents into point-like dipoles erroneously results in the hidden momentum being omitted, hence the error in [Phys. Rev. Lett. 108, 193901 (2012)]. Naturally, the Ampere-Lorentz stress tensor accords with special relativity. The Minkowski stress tensor is also consistent with special relativity. It is worth noting that the mathematical expression of the Minkowski stress tensor can be quite different from the well-known form of this stress tensor in the literature. We show that the Einstein-Laub stress tensor is incompatible with special relativity, and therefore we rebut the Einstein-Laub force density. Since the Abraham momentum density of the EM fields is inherently corresponding to the Einstein-Laub force density [Phys. Rev. Lett 111, 043602 (2013)], our rebuttal may also shed light on the controversy over the momentum of light. △ Less

Submitted 13 May, 2014; v1 submitted 1 December, 2013; originally announced December 2013.

Comments: 3 figures

Journal ref: Physical Review A 89, 043845 (2014)