Learning to perceive the world as articulated: an approach for hierarchical learning in sensory-motor systems

By Jun Tani, Stefano Nolfi. 1998

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Lifelong learning of human actions with deep neural network self-organization (2017, background): “Tani (2003) and Tani and Nolfi (1999) proposed that sensory-motor patterns for on-line planning in a robot learning scenario can be generated and recognized by using recurrent neural network models trained through prediction error minimization.”
Exploring the T-Maze: Evolving Learning-Like Robot Behaviors Using CTRNNs (2003, background): “A different line of research has studied how agents in a self-organized ways can learn internal models of the environment [9].”
Multiple Model-Based Reinforcement Learning (2002, methodology): “Tani and Nolﬁ (1999) used a similar mechanism for hierarchical sequence prediction.”
Sparse-MLP: A Fully-MLP Architecture with Conditional Computation (2021, methodology): “2017) has been successfully applied to many domains and tasks (Gavrila and Munder 2006; Hu, Palreddy, and Tompkins 1997; Tani and Nolfi 1999; Sminchisescu et al. 2004; Zeevi, Meir, and Adler 1997).”
In Search of the Neural Circuits of Intrinsic Motivation (2007, methodology): “The idea of using multiple experts has been already explored in several works including for instance (Doya et al., 2002; Baldassarre, 2002; Jordan and Jacobs, 1994; Kawato, 1999; Khamassi et al., 2005; Tani and Nolfi, 1999)”
From hardware and software to kernels and envelopes: a concept shift for robotics, developmental psychology, and brain sciences (2011, methodology): “The idea of using multiple experts has been already explored in several works including for instance (Jordan and Jacobs, 1994; Tani and Nolfi, 1999; Kawato, 1999; Doya et al., 2002; Baldassarre, 2002; Khamassi et al., 2005) The basic circuits we just described permit to compute an internal reward…”
The challenges of joint attention (2006, background): “Tani and Nolfi presented a system capable of combining local experts using gated modules (Tani and Nolfi, 1999).”
Emergence of Functional Hierarchy in a Multiple Timescale Neural Network Model: A Humanoid Robot Experiment (2008, background): “There exist earlier studies on the computational modeling of functional hierarchy in sequences of motor primitives, representative examples being the ‘‘mixture of expert’’ model [16] and the ‘‘MOSAIC’’ model [17].”
Learning to generate articulated behavior through the bottom-up and the top-down interaction processes (2003, unknown): “[15] Tani, J.”
Scaling Vision with Sparse Mixture of Experts (2021, background): “MoEs have successfully used this form of conditional computation in a range of applications [23, 30, 58, 55, 67].”
Active inference and epistemic value (2015, background): “In turn, this can improve learning and inference, which depend sensitively on an efficient and sparse (hierarchical) representation of active sampling and sensorimotor learning (Ballard et al., 2013; Barto et al., 2004; Tani & Nolfi, 1999).”