Imperfect Memory Research Articles

Macroeconomic Expectations and Cognitive Noise

This paper examines forecast biases through cognitive noise, moving beyond the conventional view that frictions emerge solely from using external data. By extending Sims’s (2003) imperfect attention model to include imperfect memory, I propose a framework where cognitive constraints impact both external and internal information use. This innovation reveals horizon-dependent forecast sensitivity: short-term forecasts adjust sluggishly while long-term forecasts may overreact. I explore the macroeconomic impact of this behavior, showing how long-term expectations, heavily influenced by current economic conditions, heighten inflation volatility. Moreover, structural estimation indicates that neglecting imperfect memory critically underestimates the informational challenges forecasters encounter.

Why death and aging ? All memories are imperfect

Recent papers have emphasized the primary role of cellular information management in biological and evolutionary development. In this framework, intelligent cells collectively measure environmental cues to improve informational validity to support natural cellular engineering as collaborative decision-making and problem-solving in confrontation with environmental stresses. These collective actions are crucially dependent on cell-based memories as acquired patterns of response to environmental stressors. Notably, in a cellular self-referential framework, all biological information is ambiguous. This conditional requirement imposes a previously unexplored derivative. All cellular memories are imperfect. From this atypical background, a novel theory of aging and death is proposed. Since cellular decision-making is memory-dependent and biology is a continuous natural learning system, the accumulation of previously acquired imperfect memories eventually overwhelms the flexibility cells require to react adroitly to contemporaneous stresses to support continued cellular homeorhetic balance. The result is a gradual breakdown of the critical ability to efficiently measure environmental information and effect cell-cell communication. This age-dependent accretion governs senescence, ultimately ending in death as an organism-wide failure of cellular networking. This approach to aging and death is compatible with all prior theories. Each earlier approach illuminates different pertinent cellular signatures of this ongoing, obliged, living process.

Quantized Magnetic Domain Wall Synapse for Efficient Deep Neural Networks.

The quantization of synaptic weights using emerging nonvolatile memory (NVM) devices has emerged as a promising solution to implement computationally efficient neural networks on resource constrained hardware. However, the practical implementation of such synaptic weights is hampered by the imperfect memory characteristics, specifically the availability of limited number of quantized states and the presence of large intrinsic device variation and stochasticity involved in writing the synaptic states. This article presents on-chip training and inference of a neural network using quantized magnetic domain wall (DW)-based synaptic array and CMOS peripheral circuits. A rigorous model of the magnetic DW device considering stochasticity and process variations has been utilized for the synapse. To achieve stable quantized weights, DW pinning has been achieved by means of physical constrictions. Finally, VGG8 architecture for CIFAR-10 image classification has been simulated by using the extracted synaptic device characteristics. The performance in terms of accuracy, energy, latency, and area consumption has been evaluated while considering the process variations and nonidealities in the DW device as well as the peripheral circuits. The proposed quantized neural network (QNN) architecture achieves efficient on-chip learning with 92.4% and 90.4% training and inference accuracy, respectively. In comparison to pure CMOS-based design, it demonstrates an overall improvement in area, energy, and latency by 13.8 × , 9.6 × , and 3.5 × , respectively.

Selfish Corporations

Abstract We study how perceptions of corporate responsibility influence policy preferences and the effectiveness of corporate communication when agents have imperfect memory recall. Using a new large-scale survey of U.S. citizens on their support for corporate bailouts, we first establish that the public demands corporations to behave better within society, a sentiment we label “big business discontent.” Using random variation in the order of survey sections and in the exposure to animated videos, we then show that priming respondents to think about corporate responsibility lowers the support for bailouts. This finding suggests that big business discontent influences policy preferences. Furthermore, we find that messages which paint a positive picture of corporate responsibility can “backfire,” as doing so brings attention to an aspect on which the public has negative views. In contrast, reframing corporate bailouts in terms of economic tradeoffs increases support for the policy. We develop a memory-based model of decision-making and communication to rationalize these findings.

Exact rate analysis for quantum repeaters with imperfect memories and entanglement swapping as soon as possible

Exact rate analysis for quantum repeaters with imperfect memories and entanglement swapping as soon as possible

Communicability in time-varying networks with memory

We develop a first-principles approach to define the communicability between two nodes in a time-varying network with memory. The formulation is based on the time-fractional Schrödinger equation, where the fractional (of Caputo type) derivative accounts for the memory of the system. Using a time-varying Hamiltonian in the tight-binding formalism we propose the temporal communicability as the product of Mittag–Leffer functions of the adjacency matrices of the temporal snapshots. We then show that the resolvent- and exponential-communicabilities of a network are special cases of the proposed temporal communicability when perfect (resolvent) or imperfect (exponential) memory are considered for the system. By using theoretical and empirical evidence we show that real-world systems work out of perfect memory, and with an interrelation between memory-dependent temporal communication and imperfect memory spatial transmission of information. We illustrate our results with the study of trophallaxis interactions in two ant colonies.

Reinforcement with Fading Memories

We study the effect of imperfect memory on decision making in the context of a stochastic sequential action-reward problem. An agent chooses a sequence of actions, which generate discrete rewards at different rates. She is allowed to make new choices at rate β, whereas past rewards disappear from her memory at rate μ. We focus on a family of decision rules where the agent makes a new choice by randomly selecting an action with a probability approximately proportional to the amount of past rewards associated with each action in her memory. We provide closed form formulas for the agent’s steady-state choice distribution in the regime where the memory span is large ([Formula: see text]) and show that the agent’s success critically depends on how quickly she updates her choices relative to the speed of memory decay. If [Formula: see text], the agent almost always chooses the best action (that is, the one with the highest reward rate). Conversely, if [Formula: see text], the agent chooses an action with a probability roughly proportional to its reward rate.

The effect of imperfect memory recall on risk preferences

AbstractPeople often take on different levels of risk when deciding on future actions compared with when they take actions with immediate consequences. The presently reported research investigated how imperfect recall of previously chosen options influences temporal fluctuations in risk preferences. In two incentivized studies with a waiting time before decision resolution (a 4‐week waiting period from an initial decision), we observed that errors in the recall of previous choices played a substantial role in constructing risk preferences for choices that are usually resolved after a delay. More specifically, in a first study, we found that after a waiting period (i.e., after participants waited for their risky choices to be resolved), participants perceived that they took less risk during their initial choices than they actually did. Importantly, when asked about whether they would change their initial choices, participants declared that they would take riskier choices. In a second study, we tested directly whether such a shift in initial choices could also be demonstrated behaviorally. Results revealed that after beliefs about past choices were induced, participants changed their risk preferences in line with the information presented during belief induction.

On the category adjustment model: another look at Huttenlocher, Hedges, and Vevea (2000)

Huttenlocher, Hedges, and Vevea (2000) (Why do categories affect stimulus judgment? Journal of Experimental Psychology: General, 129, 220-241) introduce the category adjustment model (CAM). Given that participants imperfectly remember stimuli (which we refer to as “targets”), CAM holds that participants maximize accuracy by using information about the distribution of the targets to improve their judgments. CAM predicts that judgments will be a weighted average of the imperfect memory of the target and the mean of the distribution of targets. Huttenlocher, Hedges, and Vevea (2000) report on three experiments and conclude that CAM is “verified.” We attempt to replicate Experiment 3 from Huttenlocher et al. (2000). We analyze judgment-level data rather than averaged data. We find evidence of a bias toward a set of recent targets rather than a bias toward the running mean. We also do not find evidence of the joint hypothesis that participants learned the distribution of targets and employed this information in their judgments. The judgments in our dataset are not consistent with CAM. We discuss how the apparent defects in HHV – including dividing by zero – went unnoticed and how such mistakes can be avoided in future research. Finally, we hope that the techniques that we employ will be used to test other datasets that are currently regarded as consistent with CAM or any Bayesian model of judgment.

Does Truth Have a Future?

Does Truth Have a Future?

Reinforcement with Fading Memories

We study the effect of imperfect memory on decision making in the context of a stochastic sequential action-reward problem. An agent chooses a sequence of actions which generate discrete rewards at different rates. She is allowed to make new choices at rate $\beta$, while past rewards disappear from her memory at rate $\mu$. We focus on a family of decision rules where the agent makes a new choice by randomly selecting an action with a probability approximately proportional to the amount of past rewards associated with each action in her memory. We provide closed-form formulae for the agent's steady-state choice distribution in the regime where the memory span is large ($\mu \to 0$), and show that the agent's success critically depends on how quickly she updates her choices relative to the speed of memory decay. If $\beta \gg \mu$, the agent almost always chooses the best action, i.e., the one with the highest reward rate. Conversely, if $\beta \ll \mu$, the agent chooses an action with a probability roughly proportional to its reward rate.

Non-linearity in gaze holding: Experimental results and possible mechanisms.

Cerebellar impairment may cause deficits in horizontal gaze holding, leading to centrifugal gaze-evoked nystagmus during fixation of eccentric targets. Healthy individuals show a weak drift leading to physiological nystagmus only at large gaze angles. These drifts are due to imperfect memory of the neural circuitry generating the eye position signals by integration of velocity signals. The cerebellum plays a crucial role in reducing the "leakiness" of this neural integrator. This neural integrator has been traditionally modeled as a first order low-pass filter, implying a linear relation between drift velocity and eye eccentricity. Evidences of a non-linear behavior, however, can be found in the literature. In a recent series of papers we showed that the eye drift velocity (V) can be descriptively modeled as a tangent function of gaze eccentricity (P) with the following equation: V=k2/k1tan(k1P). Notably, the two parameters have distinct roles: k1 regulates the rate of compression of the tangent, exclusively determining the non-linearity; k2 is a pure scaling factor. This descriptive model robustly fitted the data of healthy individuals both at baseline (n=50) and following transient cerebellar impairment induced by controlled amounts of alcohol [blood alcohol content 0.06% (n=15) and 0.1% (n=15)] and of patients with chronic cerebellar impairment of various origin (n=20). Interestingly, alcohol selectively changed the scaling factor k2, evidencing that an overall, transient cerebellar impairment does not impact the non-linear behavior. Patients with cerebellar degeneration, on the other hand, showed a change in both k1 and k2, implying a role of the cerebellum in limiting the range of eye positions where the non-linearity becomes relevant. Non-linearity has been reported in literature for both the neural integrator and the eye plant. While previous models using a neural network attempted to reproduce the non-linear behavior of the brainstem, we propose a block diagram reproducing the observed PV tangent relation by introducing a position dependency in the parameters of the cerebellar feedback loop.

Reinforcement with Fading Memories

We study the effect of imperfect memory on decision making in the context of a stochastic sequential action-reward problem. An agent chooses a sequence of actions which generate discrete rewards at different rates. She is allowed to make new choices at rate β, while past rewards disappear from her memory at rate μ. We focus on a family of decision rules where the agent makes a new choice by randomly selecting an action with a probability approximately proportional to the amount of past rewards associated with each action in her memory. We provide closed-form formulae for the agent's steady-state choice distribution in the regime where the memory span is large (μ -> 0), and show that the agent's success critically depends on how quickly she updates her choices relative to the speed of memory decay. If β >> μ, the agent almost always chooses the best action, i.e., the one with the highest reward rate. Conversely, if β << μ, the agent chooses an action with a probability roughly proportional to its reward rate.

An experimental study on cumulative prospect theory learning model of travelers’ dynamic mode choice under uncertainty

Abstract In this paper, we examined travelers’ dynamic mode choice behavior under travel time variability. We found travelers’ inconsistent risk attitudes through a binary mode choice experiment. Although the results deviated from the traditional utility maximization theory and could not be explained by the payoff variability effect, they could be well captured in a cumulative prospect theory (CPT) framework. After considering the imperfect memory effect, we found that the prediction ability of the cumulative prospect theory learning (CPTL) model could be significantly improved. The experimental results were also compared with the CPTL model and the reinforcement learning (REL) model. This study empirically showed the potential of alternative theories to better capture travelers’ day-to-day mode choice behavior under uncertainty. A new definition of willingness to pay (WTP) in a CPT framework was provided to explicitly consider travelers’ perceived value increases in travel time.

Robotic autonomous behavior selection using episodic memory and attention system

PurposeThis paper aims to focus on the autonomous behavior selection issue of robotics from the perspective of episodic memory in cognitive neuroscience with biology-inspired attention system. It instructs a robot to follow a sequence of behaviors. This is similar to human travel to a target location by guidance.Design/methodology/approachThe episodic memory-driving Markov decision process is proposed to simulate the organization of episodic memory by introducing neuron stimulation mechanism. Based on the learned episodic memory, the robotic global planning method is proposed for efficient behaviors sequence prediction using bottom-up attention. Local behavior planning based on risk function and feasible paths is used for behavior reasoning under imperfect memory. Aiming at the problem of whole target selection under redundant environmental information, a top-down attention servo control method is proposed to effectively detect the target containing multi-parts and distractors which share same features with the target.FindingsBased on the proposed method, the robot is able to accumulate experience through memory, and achieve adaptive behavior planning, prediction and reasoning between tasks, environment and threats. Experimental results show that the method can balance the task objectives, select the suitable behavior according to current environment.Originality/valueThe behavior selection method is integrated with cognitive levels to generate optimal behavioral sequence. The challenges in robotic planning under uncertainty and the issue of target selection under redundant environment are addressed.

On the Interaction of Memory and Procrastination: Implications for Reminders, Deadlines, and Empirical Estimation

The interaction between present bias and limited memory can explain why individuals do not act at deadlines and why providing reminders can have large effects. Individuals in my model must choose when and whether to complete a task, but may forget or procrastinate. A calibration exercise shows that assuming perfect memory leads to biased estimates of present bias because the rate of task completion at the deadline is much lower with imperfect memory. Naive procrastination explains why individuals do not set up reminders despite large gains to doing so. The model offers guidance for empirical studies of reminders, making a distinction between anticipated and unanticipated reminders: anticipated reminders can induce additional procrastination, lowering both welfare and the probability the task is completed. I then use this framework to show how to optimally set deadlines and time the delivery of reminders to present-biased individuals.

Episodic Memory-Based Robotic Planning Under Uncertainty

This paper presents a robotic behavior planning method under uncertainty based on biology-inspired episodic memory. Adaptive behavior planning, prediction and reasoning are achieved between tasks, environment, and threats. Through building a novel episode model and introducing the activation and stimulation mechanism of state neurons, the framework of an episodic memory-driving Markov decision process (EM-MDP) is proposed for incremental self-learning of robotic experience and cognitive behavior planning. Two main challenges in robot behavior control under uncertainty are addressed: high computational complexity and perceptual aliasing. The approach for robotic global planning and behaviors sequence prediction based on the EM-MDP is developed utilizing neuron synaptic potential. A local behavioral planning method based on risk function and feasible paths is employed to achieve path optimization and behavior reasoning under the condition of imperfect memory. Robot can evaluate the past events sequence, predict the current state, and plan the desired behaviors. The proposed method is evaluated in several real-life environments for a mobile robot system. The robot system is able to successfully produce solutions in general scenarios under uncertainty.

On the Category Adjustment Model: Another Look at Huttenlocher, Hedges, and Vevea (2000)

Huttenlocher, Hedges, and Vevea (2000) (Why do categories affect stimulus judgment? Journal of Experimental Psychology: General, 129, 220-241) introduce the category adjustment model (CAM). Given that participants imperfectly remember stimuli (which we refer to as “targets”), CAM holds that participants maximize accuracy by using information about the distribution of the targets to improve their judgments. CAM predicts that judgments will be a weighted average of the imperfect memory of the target and the mean of the distribution of targets. Huttenlocher, Hedges, and Vevea (2000) report on three experiments and conclude that CAM is “verified.” We attempt to replicate Experiment 3 from Huttenlocher et al. (2000). We analyze judgment-level data rather than averaged data. We find evidence of a bias toward a set of recent targets rather than a bias toward the running mean. We also do not find evidence of the joint hypothesis that participants learned the distribution of targets and employed this information in their judgments. The judgments in our dataset are not consistent with CAM. We discuss how the apparent defects in HHV – including dividing by zero – went unnoticed and how such mistakes can be avoided in future research. Finally, we hope that the techniques that we employ will be used to test other datasets that are currently regarded as consistent with CAM or any Bayesian model of judgment.

Opportunistic Channel Selection by Cognitive Wireless Nodes Under Imperfect Observations and Limited Memory: A Repeated Game Model

We study the problem of how autonomous cognitive nodes (CNs) can arrive at an efficient and fair opportunistic channel access policy in scenarios where channels may be non-homogeneous in terms of primary user (PU) occupancy. In our model, a CN that is able to adapt to the environment is limited in two ways. First, CNs have imperfect observations (such as due to sensing and channel errors) of their environment. Second, CNs have imperfect memory due to limitations in computational capabilities. For efficient opportunistic channel access, we propose a simple adaptive win-shift lose-randomize (WSLR) strategy that can be executed by a two-state machine (automaton). Using the framework of repeated games (with imperfect observations and limited memory), we show that the proposed strategy enables the CNs (without any explicit coordination) to reach an outcome that: 1) maximizes the total network payoff and also ensures fairness among the CNs; 2) reduces the likelihood of collisions among CNs; and 3) requires a small number of sensing steps (attempts) to find a channel free of PU activity. We compare the performance of the proposed autonomous strategy with a centralized strategy and also test it with real spectrum data collected at RWTH Aachen.

A Bounded Rationality Model of Information Search and Choice in Preference Measurement

It is becoming increasingly easier for researchers and practitioners to collect eye-tracking data during online preference measurement tasks. The authors develop a dynamic discrete choice model of information search and choice under bounded rationality, which they calibrate using a combination of eye-tracking and choice data. Their model extends Gabaix et al.'s (2006) directed cognition model by capturing fatigue, proximity effects, and imperfect memory encoding and by estimating individual-level parameters and partworths within a likelihood-based hierarchical Bayesian framework. The authors show that modeling eye movements as the outcome of forward-looking utility maximization improves out-of-sample predictions, enables researchers and practitioners to use shorter questionnaires, and allows better discrimination between attributes.