Many current neurophysiological, psychophysical, and psychological approaches to vision rest on the idea that when we see, the brain produces an internal representation of the world. The activation of this internal representation is assumed to give rise to the experience of seeing. The problem with this kind of approach is that it leaves unexplained how the existence of such a detailed internal representation might produce visual consciousness. An alternative proposal is made here. We propose that seeing is a way of acting. It is a particular way of exploring the environment. Activity in internal representations does not generate the experience of seeing. The outside world serves as its own, external, representation. The experience of seeing occurs when the organism masters what we call the governing laws of sensorimotor contingency. The advantage of this approach is that it provides a natural and principled way of accounting for visual consciousness, and for the differences in the perceived quality of sensory experience in the different sensory modalities. Several lines of empirical evidence are brought forward in support of the theory, in particular: evidence from experiments in sensorimotor adaptation, visual \"filling in,\" visual stability despite eye movements, change blindness, sensory substitution, and color perception.

A sensorimotor account of vision and visual consciousness-Authors' Response-Acting out our sensory experience

Model-free reinforcement learning (RL) can be used to learn effective policies for complex tasks, such as Atari games, even from image observations. However, this typically requires very large amounts of interaction -- substantially more, in fact, than a human would need to learn the same games. How can people learn so quickly? Part of the answer may be that people can learn how the game works and predict which actions will lead to desirable outcomes. In this paper, we explore how video prediction models can similarly enable agents to solve Atari games with fewer interactions than model-free methods. We describe Simulated Policy Learning (SimPLe), a complete model-based deep RL algorithm based on video prediction models and present a comparison of several model architectures, including a novel architecture that yields the best results in our setting. Our experiments evaluate SimPLe on a range of Atari games in low data regime of 100k interactions between the agent and the environment, which corresponds to two hours of real-time play. In most games SimPLe outperforms state-of-the-art model-free algorithms, in some games by over an order of magnitude.

Model-Based Reinforcement Learning for Atari

/pdf/model-based-reinforcement-learning-for-atari-3xtcq1mkpb.pdf

Model Based Reinforcement Learning for Atari

A review on EMG-based motor intention prediction of continuous human upper limb motion for human-robot collaboration

During the last few years, Convolutional Neural Networks are slowly but surely becoming the default method solve many computer vision related problems. This is mainly due to the continuous success that they have achieved when applied to certain tasks such as image, speech, or object recognition.

PointNet: A 3D Convolutional Neural Network for real-time object class recognition

Service robots are expected to play an important role in our daily lives as our companions in home and work environments in the near future. An important requirement for fulfilling this expectation is to equip robots with skills to perform everyday manipulation tasks, the success of which is crucial for most home chores, such as cooking, cleaning, and shopping. Robots have been used successfully for manipulation tasks in wellstructured and controlled factory environments for decades. Designing skills for robots working in uncontrolled human environments raises many potential challenges in various subdisciplines, such as computer vision, automated planning, and human-robot interaction. In spite of the recent progress in these fields, there are still challenges to tackle. This article outlines problems in different research areas related to mobile manipulation from the cognitive perspective, reviews recently published works and the state-of-the-art approaches to address these problems, and discusses open problems to be solved to realize robot assistants that can be used in manipulation tasks in unstructured human environments.

Cognition-Enabled Robot Manipulation in Human Environments: Requirements, Recent Work, and Open Problems

We propose a visual scene interpretation system for cognitive robots to maintain a consistent world model about their environments. This interpretation system is for our lifelong experimental learning framework that allows robots analyze failure contexts to ensure robustness in their future tasks. To efficiently analyze failure contexts, scenes should be interpreted appropriately. In our system, LINE-MOD and HS histograms are used to recognize objects with/without textures. Moreover, depth-based segmentation is applied for identifying unknown objects in the scene. This information is also used to augment the recognition performance. The world model includes not only the objects detected in the environment but also their spatial relations to efficiently represent contexts. Extracting unary and binary relations such as on, on_ground, clear and near is useful for symbolic representation of the scenes. We test the performance of our system on recognizing objects, determining spatial predicates, and maintaining consistency of the world model of the robot in the real world. Our preliminary results reveal that our system can be successfully used to extract spatial relations in a scene and to create a consistent model of the world by using the information gathered from the onboard RGB-D sensor as the robot explores its environment.

https://www.aaai.org/ocs/index.php/SSS/SSS14/paper/download/7762/7704

Scene Interpretation for Self-Aware Cognitive Robots

Service robots need to handle a variety of everyday manipulation tasks to accomplish household chores such as cooking and cleaning. Successful execution of these tasks is highly dependent on how reliable the robot perceives its environment through noisy sensing. We present a visual world modeling system for service robots to generate and maintain accurate models of their environments for continuous scenarios. This system is designed to provide a generic platform for both humanoid and ground manipulation robots using different types of vision sensors and algorithms. In our particular implementation, visual data processed by different perception algorithms are used for building and continuously updating a world model of the environment. We evaluate our system on a variety of object manipulation scenarios and show that the system produces consistent perception outcomes suitable for different manipulation tasks.

Continuous Visual World Modeling for Autonomous Robot Manipulation

A cognitive robot may face failures during the execution of its actions. These failures are mostly due to the gap between the phys- ical world and the constructed symbolic plans, some internal problems that may occur in its embodiment or unexpected external events. In this paper, we propose a visual scene interpretation system for extracting spatial relations among objects in a scene and using these relations to detect failures during the plan execution. Our system uses LINE-MOD and HS histograms in order to recognize textureless objects with different shapes and colors. Then, it analyzes the scene to specify the world state after each action execution. Our focus in this research is on particularly the following spatial relations: on, on table, clear and unstable. In the experiments, we test the performance of our system on recognizing ob- jects, determining pairwise spatial relations among them, and detecting failures using these relations. Our preliminary results reveal that our sys- tem can be successfully used to extract spatial relations in a scene, and to determine failures during plan execution by using this information.

Extracting Spatial Relations Among Objects for Failure Detection

A cognitive robot may face failures during the execution of its actions in the physical world. In this paper, we investigate how robots can ensure robustness by gaining experience on action executions, and we propose a lifelong experimental learning method. We use Inductive Logic Programming (ILP) as the learning method to frame new hypotheses. ILP provides first-order logic representations of the derived hypotheses that are useful for reasoning and planning processes. Furthermore, it can use background knowledge to represent more advanced rules. Partially specified world states can also be easily represented in these rules. All these advantages of ILP make this approach superior to attribute-based learning approaches. Experience gained through incremental learning is used as a guide to future decisions of the robot for robust execution. The results on our Pioneer 3DX robot reveal that the hypotheses framed for failure cases are sound and ensure safety in future tasks of the robot.

/pdf/learning-guided-planning-for-robust-task-execution-in-1kgps6wmwq.pdf

Mustafa Ersen

Papers

Cognition-Enabled Robot Manipulation in Human Environments: Requirements, Recent Work, and Open Problems

Scene Interpretation for Self-Aware Cognitive Robots

Continuous Visual World Modeling for Autonomous Robot Manipulation

Extracting Spatial Relations Among Objects for Failure Detection

Learning Guided Planning for Robust Task Execution in Cognitive Robotics