scholarly journals Optimal policy for multi-alternative decisions

2019 ◽  
Author(s):  
Satohiro Tajima ◽  
Jan Drugowitsch ◽  
Nisheet Patel ◽  
Alexandre Pouget
Keyword(s):  
Optimal Policy ◽  
Neural Circuit ◽  
Internal Variability ◽  
Time Dependent ◽  
Irrelevant Alternative ◽  
Evidence Accumulation ◽  
Divisive Normalization ◽  
Key Features ◽  
Similarity Effect ◽  
Alternative Principle

AbstractEvery-day decisions frequently require choosing among multiple alternatives. Yet, the optimal policy for such decisions is unknown. Here we derive the normative policy for general multi-alternative decisions. This strategy requires evidence accumulation to nonlinear, time-dependent bounds, that trigger choices. A geometric symmetry in those boundaries allows the optimal strategy to be implemented by a simple neural circuit involving a normalization with fixed decision bounds and an urgency signal. The model captures several key features of the response of decision-making neurons as well as the increase in reaction time as a function of the number of alternatives, known as Hick’s law. In addition, we show that, in the presence of divisive normalization and internal variability, our model can account for several so called ‘irrational’ behaviors such as the similarity effect as well as the violation of both the independent irrelevant alternative principle and the regularity principle.

scholarly journals The Interpretation and Use of Biases in Decadal Climate Predictions

2014 ◽  
Vol 27 (8) ◽  
pp. 2931-2947 ◽  
Author(s):  
Ed Hawkins ◽  
Buwen Dong ◽  
Jon Robson ◽  
Rowan Sutton ◽  
Doug Smith
Keyword(s):  
Climate Model ◽  
Coupled Model ◽  
Internal Variability ◽  
Time Dependent ◽  
Start Time ◽  
Transient Climate Response ◽  
Global Mean Temperature ◽  
Decadal Climate ◽  
Highly Correlated ◽  
Hadley Centre

Abstract Decadal climate predictions exhibit large biases, which are often subtracted and forgotten. However, understanding the causes of bias is essential to guide efforts to improve prediction systems, and may offer additional benefits. Here the origins of biases in decadal predictions are investigated, including whether analysis of these biases might provide useful information. The focus is especially on the lead-time-dependent bias tendency. A “toy” model of a prediction system is initially developed and used to show that there are several distinct contributions to bias tendency. Contributions from sampling of internal variability and a start-time-dependent forcing bias can be estimated and removed to obtain a much improved estimate of the true bias tendency, which can provide information about errors in the underlying model and/or errors in the specification of forcings. It is argued that the true bias tendency, not the total bias tendency, should be used to adjust decadal forecasts. The methods developed are applied to decadal hindcasts of global mean temperature made using the Hadley Centre Coupled Model, version 3 (HadCM3), climate model, and it is found that this model exhibits a small positive bias tendency in the ensemble mean. When considering different model versions, it is shown that the true bias tendency is very highly correlated with both the transient climate response (TCR) and non–greenhouse gas forcing trends, and can therefore be used to obtain observationally constrained estimates of these relevant physical quantities.

A Canonical Neural Circuit for Cortical Nonlinear Operations

2008 ◽  
Vol 20 (6) ◽  
pp. 1427-1451 ◽  
Author(s):  
Minjoon Kouh ◽  
Tomaso Poggio
Keyword(s):  
Neural Circuit ◽  
Neural Responses ◽  
Energy Models ◽  
Divisive Normalization ◽  
Primate Visual Cortex ◽  
Neurophysiological Data ◽  
Ventral Pathway ◽  
Response Profiles ◽  
Parameter Values ◽  
Good Agreement

A few distinct cortical operations have been postulated over the past few years, suggested by experimental data on nonlinear neural response across different areas in the cortex. Among these, the energy model proposes the summation of quadrature pairs following a squaring nonlinearity in order to explain phase invariance of complex V1 cells. The divisive normalization model assumes a gain-controlling, divisive inhibition to explain sigmoid-like response profiles within a pool of neurons. A gaussian-like operation hypothesizes a bell-shaped response tuned to a specific, optimal pattern of activation of the presynaptic inputs. A max-like operation assumes the selection and transmission of the most active response among a set of neural inputs. We propose that these distinct neural operations can be computed by the same canonical circuitry, involving divisive normalization and polynomial nonlinearities, for different parameter values within the circuit. Hence, this canonical circuit may provide a unifying framework for several circuit models, such as the divisive normalization and the energy models. As a case in point, we consider a feedforward hierarchical model of the ventral pathway of the primate visual cortex, which is built on a combination of the gaussian-like and max-like operations. We show that when the two operations are approximated by the circuit proposed here, the model is capable of generating selective and invariant neural responses and performing object recognition, in good agreement with neurophysiological data.

scholarly journals An optimal policy for deteriorating items with time-proportional deterioration rate and constant and time-dependent linear demand rate

2017 ◽  
Vol 13 (4) ◽  
pp. 455-463 ◽  
Author(s):  
Trailokyanath Singh ◽  
Pandit Jagatananda Mishra ◽  
Hadibandhu Pattanayak
Keyword(s):  
Optimal Policy ◽  
Deteriorating Items ◽  
Time Dependent ◽  
Deterioration Rate ◽  
Demand Rate ◽  
Linear Demand

scholarly journals Internal Variability of the Winter Stratosphere. Part II: Time-Dependent Forcing

2008 ◽  
Vol 65 (7) ◽  
pp. 2375-2388 ◽  
Author(s):  
R. K. Scott ◽  
L. M. Polvani ◽  
D. W. Waugh
Keyword(s):  
Lower Boundary ◽  
Polar Vortex ◽  
Internal Variability ◽  
Time Dependent ◽  
Random Wave ◽  
Frequency Locking ◽  
Wave Forcing ◽  
Wave Flux ◽  
Time Periodic ◽  
Zero Time

Abstract This paper considers the effect of time-dependent lower boundary wave forcing on the internal variability found to appear spontaneously in a stratosphere-only model when the forcing is perfectly steady. While the time-dependent forcing is found to modulate the internal variability, leading in some cases to frequency locking of the upper-stratospheric response to the forcing, the temporal and spatial structure of the variability remains similar to the case when the forcing is time independent. Experiments with a time-periodic modulation of the forcing amplitude indicate that the wave flux through the lower boundary is only partially related to the instantaneous forcing, but is more significantly influenced by the condition of the polar vortex itself. In cases of purely random wave forcing with zero time mean, the stratospheric response is similar to that obtained with steady forcing of magnitude equal to the root-mean-square of the time-varying forcing.

scholarly journals Where’s the noise? Key features of neuronal variability and inference emerge from self-organized learning

2014 ◽  
Author(s):  
Christoph Hartmann ◽  
Andreea Lazar ◽  
Jochen Triesch
Keyword(s):  
Spontaneous Activity ◽  
Network Model ◽  
Neural Circuit ◽  
Intrinsic Noise ◽  
Spike Timing ◽  
Homeostatic Plasticity ◽  
Neural Recordings ◽  
Self Organized ◽  
Key Features ◽  
Neural Variability

AbstractTrial-to-trial variability and spontaneous activity of cortical recordings have been suggested to reflect intrinsic noise. This view is currently challenged by mounting evidence for structure in these phenomena: Trial-to-trial variability decreases following stimulus onset and can be predicted by previous spontaneous activity. This spontaneous activity is similar in magnitude and structure to evoked activity and can predict decisions. Allof the observed neuronal properties described above can be accounted for, at an abstract computational level, by the sampling-hypothesis, according to which response variability reflects stimulus uncertainty. However, a mechanistic explanation at the level of neural circuit dynamics is still missing.In this study, we demonstrate that all of these phenomena can be accounted for by a noise-free self-organizing recurrent neural network model (SORN). It combines spike-timing dependent plasticity (STDP) and homeostatic mechanisms in a deterministic network of excitatory and inhibitory McCulloch-Pitts neurons. The network self-organizes to spatio-temporally varying input sequences.We find that the key properties of neural variability mentioned above develop in this model as the network learns to perform sampling-like inference. Importantly, the model shows high trial-to-trial variability although it is fully deterministic. This suggests that the trial-to-trial variability in neural recordings may not reflect intrinsic noise. Rather, it may reflect a deterministic approximation of sampling-like learning and inference. The simplicity of the model suggests that these correlates of the sampling theory are canonical properties of recurrent networks that learn with a combination of STDP and homeostatic plasticity mechanisms.Author SummaryNeural recordings seem very noisy. If the exact same stimulus is shown to an animal multiple times, the neural response will vary. In fact, the activity of a single neuron shows many features of a stochastic process. Furthermore, in the absence of a sensory stimulus, cortical spontaneous activity has a magnitude comparable to the activity observed during stimulus presentation. These findings have led to a widespread belief that neural activity is indeed very noisy. However, recent evidence indicates that individual neurons can operate very reliably and that the spontaneous activity in the brain is highly structured, suggesting that much of the noise may in fact be signal. One hypothesis regarding this putative signal is that it reflects a form of probabilistic inference through sampling. Here we show that the key features of neural variability can be accounted for in a completely deterministic network model through self-organization. As the network learns a model of its sensory inputs, the deterministic dynamics give rise to sampling-like inference. Our findings show that the notorious variability in neural recordings does not need to be seen as evidence for a noisy brain. Instead it may reflect sampling-like inference emerging from a self-organized learning process.

scholarly journals Effects of short-term plasticity in early olfactory information processing in Drosophila

2021 ◽  
Author(s):  
Yuxuan Liu ◽  
Qianyi Li ◽  
Chao Tang ◽  
Shanshan Qin ◽  
Yuhai Tu
Keyword(s):  
Experimental Data ◽  
Adaptive Response ◽  
Neural Circuit ◽  
Dominant Factor ◽  
Projection Neurons ◽  
Short Term ◽  
Information Process ◽  
Short Term Plasticity ◽  
Olfactory Information ◽  
Divisive Normalization

In Drosophila, olfactory information received by the olfactory receptor neurons (ORNs) is first processed by an incoherent feed forward neural circuit in the antennal lobe (AL) that consists of ORNs (input), the inhibitory local neurons (LNs), and projection neurons (PNs). This "early" olfactory information process has two important characteristics. First, response of a PN to its cognate ORN is normalized by the overall activity of other ORNs, a phenomenon termed "divisive normalization". Second, PNs respond strongly to the onset of ORN activities, but they adapt to prolonged or continuously increasing inputs. Despite the importance of these characteristics for learning and memory, their underlying mechanism remains not fully understood. Here, we develop a circuit model for describing the ORN-LN-PN dynamics by including key features of neuron-neuron interactions, in particular short-term plasticity (STP) and presynaptic inhibition (PI).Our model shows that STP is critical in shaping PN's steady-state response properties. By fitting our model to experimental data quantitatively, we found that strong and balanced short-term facilitation (STF) and short-term depression (STD) in STP is crucial for the observed nonlinear divisive normalization in Drosophila. By comparing our model with the observed adaptive response to time-varying signals quantitatively, we find that both STP and PI contribute to the highly adaptive response with the latter being the dominant factor for a better fit with experimental data. Our model not only helps reveal the mechanisms underlying two main characteristics of the early olfactory process, it can also be used to predict the PN responses to arbitrary time-dependent signals and to infer microscopic properties of the circuit (such as the strengths of STF and STD) from the measured input-output relation.

Computerized Reconstruction of Building Fires

1991 ◽  
Author(s):  
David G. Lilley
Keyword(s):  
Large Scale ◽  
Time Dependent ◽  
Behavioral Interactions ◽  
Toxic Gases ◽  
Smoke Layer ◽  
Fire Scenarios ◽  
Clean Air ◽  
Key Features ◽  
Witness Statements ◽  
Important Time

Abstract A fire development simulation model is described which provides estimates of the amount and temperature of the smoke layer produced, the evolution of toxic gases, and the amount of time available from the onset of fire for the safe departure of occupants. Its results can be used to determine the key features of the fire evolution and the corresponding danger to occupants. Studies of this type help to validate or deny the suggested fire scenario and witness statements. Mathematical modeling thus helps to discriminate between alternative fire scenarios by evaluating the consequences and comparing them with observations. The software consists of data, procedures, and computer programs which simulate important time-dependent phenomena involved in residential fires. Based on sound scientific and mathematical principles, predictions are made of the production of energy and mass (smoke and gases) by one or more burning objects in one room, based on small or large scale measurements. The buoyancy-driven transport of this energy and mass through a series of user-specified rooms and connections is then computed (doors, windows, cracks, etc.). The resulting temperatures, smoke optical densities, and gas concentrations (after accounting for heat transfer to surfaces and dilution by mixing with clean air) are linked to the problem of egress. The evacuation process of a set of occupants may be simulated, accounting for delays in notification, decision making, behavioral interactions, and inherent capabilities.

scholarly journals Visual evidence accumulation guides decision-making in unrestrained mice

2017 ◽  
Author(s):  
Onyekachi Odoemene ◽  
Sashank Pisupati ◽  
Hien Nguyen ◽  
Anne K. Churchland
Keyword(s):  
Decision Making ◽  
Neural Activity ◽  
Neural Circuits ◽  
Neural Circuit ◽  
Deep Understanding ◽  
Neural Mechanism ◽  
Neural Structure ◽  
Evidence Accumulation ◽  
Visual Evidence ◽  
Decision Making Behavior

AbstractThe ability to manipulate neural activity with precision is an asset in uncovering neural circuits for decision-making. Diverse tools for manipulating neurons are available for mice, but the feasibility of mice for decision-making studies remains unclear, especially when decisions require accumulating visual evidence. For example, whether mice’ decisions reflect leaky accumulation is not established, and the relevant and irrelevant factors that influence decisions are unknown. Further, causal circuits for visual evidence accumulation have not been established. To address these issues, we measured >500,000 decisions in 27 mice trained to judge the fluctuating rate of a sequence of flashes. Information throughout the 1000ms trial influenced choice, but early information was most influential. This suggests that information persists in neural circuits for ~1000ms with minimal accumulation leak. Further, while animals primarily based decisions on current stimulus rate, they were unable to entirely suppress additional factors: total stimulus brightness and the previous trial’s outcome. Next, we optogenetically inhibited anteromedial (AM) visual area using JAWS. Importantly, light activation biased choices in both injected and uninjected animals, demonstrating that light alone influences behavior. By varying stimulus-response contingency while holding stimulated hemisphere constant, we surmounted this obstacle to demonstrate that AM suppression biases decisions. By leveraging a large dataset to quantitatively characterize decision-making behavior, we establish mice as suitable for neural circuit manipulation studies, including the one here. Further, by demonstrating that mice accumulate visual evidence, we demonstrate that this strategy for reducing uncertainty in decision-making is employed by animals with diverse visual systems.Significance statementTo connect behaviors to their underlying neural mechanism, a deep understanding of the behavioral strategy is needed. This understanding is incomplete in mouse studies, in part because existing datasets have been too small to quantitatively characterize decision-making behavior. To surmount this, we measured the outcome of over 500,000 decisions made by 27 mice trained to judge visual stimuli. Our analyses offer new insights into mice’ decision-making strategies and compares them with those of other species. We then disrupted neural activity in a candidate neural structure and examined the effect on decisions. Our findings establish mice as a suitable organism for visual accumulation of evidence decisions. Further, the results highlight similarities in decision-making strategies across very different species.

The power of the little ones: Computed and observed aftershock hazard in Central Italy

2021 ◽  
pp. 875529302110369
Author(s):  
Robin Gee ◽  
Laura Peruzza ◽  
Marco Pagani
Keyword(s):  
Seismic Hazard ◽  
Peak Ground Acceleration ◽  
Site Effects ◽  
Hazard Analysis ◽  
Central Italy ◽  
Current Approach ◽  
Time Dependent ◽  
Ground Acceleration ◽  
Key Features ◽  
Hazard Curves

Seismic hazard in Central Italy due to the 2016–2017 seismic sequence is modeled using a standard probabilistic aftershock seismic hazard model. Two key features of the model are the consideration of time-dependent aftershock occurrence, modeled by stacking Omori decay curves associated with the three largest ( Mw > 5.5) events, and the incorporation of geologic information by modeling the locations of expected seismicity along realistic fault surfaces. The computed seismic hazard at Amatrice indicates higher hazard values compared to those computed using a conventional time-independent hazard analysis. We then compare the computed hazard curves against empirical hazard curves constructed for 12 individual recording stations in terms of peak ground acceleration, each with at least 35 (and up to 231) recordings. At eight sites, the observed exceedances fall within one standard deviation of the expected mean, while at the remaining sites, the observed exceedances fall outside this range indicating a poorer match. The soil sites are among the stations with the poorest match, suggesting that site effects may not be accurately modeled with the current approach.

scholarly journals Neural circuit mechanism for generating categorical representations

2019 ◽  
Author(s):  
Nagaraj R. Mahajan ◽  
Shreesh P. Mysore
Keyword(s):  
Decision Making ◽  
Neural Circuit ◽  
Feedback Inhibition ◽  
Neural Noise ◽  
Inhibitory Function ◽  
Recurrent Excitation ◽  
Neural Representations ◽  
Neural Inhibition ◽  
Divisive Normalization ◽  
Barn Owls

Categorical neural representations underlie various forms of selection and decision-making. They promote robust signaling of the winner in the presence of input ambiguity and neural noise. Here, we show that a ‘donut-like’ inhibitory mechanism, in which each competing option suppresses all options except itself, is highly effective at generating categorical responses. It far surpasses motifs of feedback inhibition, recurrent excitation, and divisive normalization invoked frequently in decision-making models. We demonstrate experimentally not only that this mechanism operates in the midbrain spatial selection network in barn owls, but also that it is required for categorical signaling by it. Indeed, the functional pattern of neural inhibition in the midbrain forms an exquisitely structured ‘multi-holed’ donut consistent with this network’s combinatorial inhibitory function. Moreover, simulation results reveal a generalizable neural implementation of the donut-like motif for categorization across brain areas. Self-sparing inhibition may be a powerful circuit module central to categorical selection.

Sign in / Sign up

Export Citation Format

Share Document