Explore chapters and articles related to this topic
Embodied AI, or the tale of taming the fungus eater
Published in Arkapravo Bhaumik, From AI to Robotics, 2018
The frame problem is an inquiry into determining which information is relevant to reasoning in a given situation, and which information is to be ignored. The frame problem is an issue in designing artificial situated intelligence, and the problem has a very special position in the history of AI and robots. Noted AI philosophers, Dreyfus and Searle, considered the frame problem as the demise of AI. To rescue the day, an apparent solution to the frame problem would be to associate each event with a probability. However probabilistic knowledge is hardly of much use in a real world situation and would utterly fail in hazardous situations as was the case for Dennett’s three robots. Also, modelling probability, which updates in near real-time over the robot’s sensors is computationally prohibitive [43] and at best restricted to simple worlds.
Introduction to Expert Systems
Published in Chris Nikolopoulos, Expert Systems, 1997
The frame problem is defined as the problem of finding a representational form permitting a changing, complex world to be efficiently and adequately represented, (for an analysis of the frame problem see [9]). When a change occurs in the world, the majority of things are not effected by it. How can we know which are the blocks of knowledge affected by it and which need to be updated without exhaustively checking all of the knowledge? Trying by brute force to predict all possible correlations in advance and include them in the knowledge base, clearly won't do in complex domains, as it introduces combinatorial explosion. The frame problem concerns a truly dynamic real world environment, in which circumstances that could be encountered are too vast to be modeled in a static manner. Present knowledge representation techniques involve the modeling of an idealized subset of a real world problem and suffer from their inability to adjust in unforeseen circumstances not predicted by the model. No matter which representation we choose, its effectiveness is constrained by the frame problem.
Symbolic Learning
Published in Adrian A. Hopgood, Intelligent Systems for Engineers and Scientists, 2021
The frame problem (or situation-identification problem), introduced in Chapter 1, affects many areas of artificial intelligence, particularly planning (Chapter 14). It is also pertinent here, where the problem is to determine which aspects of a given example situation are relevant to the new rule. A system for control of a boiler will have access to a wide range of information. Suppose that it comes across a set of circumstances where it is told by the teacher that it should shut off valve_2. The current world state perceived by the system is determined by stored data and sensor values, for example:
Information flow in context-dependent hierarchical Bayesian inference
Published in Journal of Experimental & Theoretical Artificial Intelligence, 2022
Chris Fields, James F. Glazebrook
The situation faced by such an observer as above is well known in AI, where it is formulated by the Frame Problem: the problem of circumscribing what does not change when an action is performed (Dietrich & Fields, 1996; McCarthy & Hayes, 1969; Shanahan, 2016). Viewed broadly, it is the problem of circumscribing what is relevant in a situation. As solving the Frame Problem requires, in effect, checking all facts in a situation to look for changes, it rapidly becomes unsolvable in practice, except approximately through the use of greedy heuristics, as the domain of action becomes large (indeed the general case is formally undecidable (Dietrich & Fields, 2020)). The problem of individual object re-identification over time can be recast as an instance of the Frame Problem (Fields, 2013b); this problem is only ‘solved’ heuristically.8
Sexbots: a case for artificial ethical agents
Published in Connection Science, 2020
Christopher James Headleand, William J. Teahan, Llyr ap Cenydd
The frame problem is the challenge of evaluating the effects of an agents action without having explicitly represent a large number of possible non-events. It described the problem of limiting the scope of reasoning by time of effect to focus on the important consequences of an action? The frame problem is one of the core challenges in artificial intelligence, and is yet to be fully addressed. Almost all normative positions fall foul of the frame problem (Wallach, 2010) as they require the agent to consider the future consequence of their actions. This requires a significant amount of information about the world which may not be available.