Input-Anticipating Critical Reservoirs Show Power Law Forgetting of Unexpected Input Events
نویسنده
چکیده
Usually reservoir computing shows an exponential memory decay. This letter investigates under which circumstances echo state networks can show a power law forgetting. That means traces of earlier events can be found in the reservoir for very long time spans. Such a setting requires critical connectivity exactly at the limit of what is permissible according to the echo state condition. However, for general matrices, the limit cannot be determined exactly from theory. In addition, the behavior of the network is strongly influenced by the input flow. Results are presented that use certain types of restricted recurrent connectivity and anticipation learning with regard to the input, where power law forgetting can indeed be achieved.
منابع مشابه
Design and Simulation of a New DC Power Supply Based on Dual Bridge Matrix Converter
A conventional high power DC power supply systems consist of a three-phase diode rectifier followed by a high frequency converter to supply loads at regulated DC voltage. These rectifiers draw significant harmonic currents from the utility, resulting in poor input power factor. In this paper, a DC power supply based on dual-bridge matrix converter (DBMC) with reduced number of switches is p...
متن کاملRobust Optimal Control of Flexible Spacecraft During Slewing Maneuvers
In this paper, slewing maneuver of a flexible spacecraft with large angle of rotation is considered and assuming structural frequency uncertainties a robust minimum-time optimal control law is developed. Considering typical bang-bang control commands with multiple symmetrical switches, a parameter optimization procedure is introduced to find the control forces/torques. The constrained minimizat...
متن کاملNeural noise can explain expansive, power-law nonlinearities in neural response functions.
Many phenomenological models of the responses of simple cells in primary visual cortex have concluded that a cell's firing rate should be given by its input raised to a power greater than one. This is known as an expansive power-law nonlinearity. However, intracellular recordings have shown that a different nonlinearity, a linear-threshold function, appears to give a good prediction of firing r...
متن کاملCOMPARISON BETWEEN MINIMUM AND NEAR MINIMUM TIME OPTIMAL CONTROL OF A FLEXIBLE SLEWING SPACECRAFT
In this paper, a minimum and near-minimum time optimal control laws are developed and compared for a rigid space platform with flexible links during an orientating maneuver with large angle of rotation. The control commands are considered as typical bang-bang with multiple symmetrical switches, the time optimal control solution for the rigid-body mode is obtained as a bang-bang function and app...
متن کاملNote on the power law of forgetting
Power functions (e.g., f(t) = at) describe the relationships among many variables observed in nature. One example of this is the power law of forgetting: The decline in memory performance with time or intervening events is well fit by a power function. This simple functional relationship accounts for a great deal of accumulated data. In this note, we consider a simple yet general memory model i...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Neural computation
دوره 27 5 شماره
صفحات -
تاریخ انتشار 2015