Using symbolic AI for knowledge-based question answering

A similar problem, called the Qualification Problem, occurs in trying to enumerate the preconditions for an action to succeed. An infinite number of pathological conditions can be imagined, e.g., a banana in a tailpipe could prevent a car from operating correctly. Japan championed Prolog for its Fifth Generation Project, intending to build special hardware for high performance. Similarly, LISP machines were built to run LISP, but as the second AI boom turned to bust these companies could not compete with new workstations that could now run LISP or Prolog natively at comparable speeds. Our chemist was Carl Djerassi, inventor of the chemical behind the birth control pill, and also one of the world’s most respected mass spectrometrists.

The early pioneers of AI believed that “every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.” Therefore, symbolic AI took center stage and became the focus of research projects. Thus contrary to pre-existing cartesian philosophy he maintained that we are born without innate ideas and knowledge is instead determined only by experience derived by a sensed perception. Children can be symbol manipulation and do addition/subtraction, but they don’t really understand what they are doing. So the ability to manipulate symbols doesn’t mean that you are thinking. A certain set of structural rules are innate to humans, independent of sensory experience. With more linguistic stimuli received in the course of psychological development, children then adopt specific syntactic rules that conform to Universal grammar.

IBM’s new AI outperforms competition in table entry search with question-answering

The advantage of neural networks is that they can deal with messy and unstructured data. Instead of manually laboring through the rules of detecting cat pixels, you can train a deep learning algorithm on many pictures of cats. The neural network then develops a statistical model for cat images.

As a result, LNNs are capable of greater understandability, tolerance to incomplete knowledge, and full logical expressivity. Figure 1 illustrates the difference between typical neurons and logical neurons. And unlike symbolic AI, neural networks have no notion of symbols and hierarchical representation of knowledge. This limitation makes it very hard to apply neural networks to tasks that require logic and reasoning, such as science and high-school math. Next, we’ve used LNNs to create a new system for knowledge-based question answering (KBQA), a task that requires reasoning to answer complex questions. Our system, called Neuro-Symbolic QA (NSQA),2 translates a given natural language question into a logical form and then uses our neuro-symbolic reasoner LNN to reason over a knowledge base to produce the answer.

Originally, researchers favored the discrete, symbolic approaches towards AI, targeting problems ranging from knowledge representation, reasoning, and planning to automated theorem proving. Integrating symbolic AI with modern machine learning techniques offers a promising path forward. This approach is particularly relevant for SEO and content marketing, where understanding and reasoning about the context of information is crucial. By leveraging symbolic reasoning, AI can enhance content discovery, improve relevance, and deliver more accurate and meaningful results, ultimately driving better engagement and conversions. In fact, rule-based AI systems are still very important in today’s applications.

The operation shown below is a variant of what is called Propositional Resolution. The expressions above the line are the premises of the rule, and the expression below is the conclusion. What distinguishes a correct pattern from one that is incorrect is that it must always lead to correct conclusions, i.e. they must be correct so long as the premises on which they are based are correct. As we will see, this is the defining criterion for what we call deduction. Obviously, there are patterns that are just plain wrong in the sense that they can lead to incorrect conclusions. Consider, as an example, the faulty reasoning pattern shown below.

And while the current success and adoption of deep learning largely overshadowed the preceding techniques, these still have some interesting capabilities to offer. In this article, we will look into some of the original symbolic AI principles and how they can be combined with deep learning to leverage the benefits of both of these, seemingly unrelated (or even contradictory), approaches to learning and AI. As illustrated in the image above, the study demonstrates that the language network in the brain is activated during language comprehension and production tasks, such as understanding or producing sentences, lists of words, and even nonwords.

From Harold Cohen to Modern AI: The Power of Symbolic Reasoning

Below is a quick overview of approaches to knowledge representation and automated reasoning. It is one form of assumption, and a strong one, while deep neural architectures contain other assumptions, usually about https://chat.openai.com/ how they should learn, rather than what conclusion they should reach. The ideal, obviously, is to choose assumptions that allow a system to learn flexibly and produce accurate decisions about their inputs.

(1) If a proposition on the left hand side of one sentence is the same as a proposition on the right hand side of the other sentence, it is okay to drop the two symbols, with the proviso that only one such pair may be dropped. (2) If a constant is repeated on the same side of a single sentence, all but one of the occurrences can be deleted. Using the methods of algebra, we can then manipulate these expressions to solve the problem.

The deep learning hope—seemingly grounded not so much in science, but in a sort of historical grudge—is that intelligent behavior will emerge purely from the confluence of massive data and deep learning. Parsing, tokenizing, spelling correction, part-of-speech tagging, noun and verb phrase chunking are all aspects of natural language processing long handled by symbolic AI, but since improved by deep learning approaches. In symbolic AI, discourse representation theory and first-order logic have been used to represent sentence meanings. Latent semantic analysis (LSA) and explicit semantic analysis also provided vector representations of documents. In the latter case, vector components are interpretable as concepts named by Wikipedia articles.

While the interest in the symbolic aspects of AI from the mainstream (deep learning) community is quite new, there has actually been a long stream of research focusing on the very topic within a rather small community called Neural-Symbolic Integration (NSI) for learning and reasoning [12]. Amongst the main advantages of this logic-based approach towards ML have been the transparency to humans, deductive reasoning, inclusion of expert knowledge, and structured generalization from small data. Read more about our work in neuro-symbolic AI from the MIT-IBM Watson AI Lab. Our researchers are working to usher in a new era of AI where machines can learn more like the way humans do, by connecting words with images and mastering abstract concepts. This dissociation seems to indicate that language is not necessary for thought.

Automated reasoning programs can be used to check proofs and, in some cases, to produce proofs or portions of proofs. The example also introduces one of the most important operations in what is symbolic reasoning Formal Logic, viz. Resolution has the property of being complete for an important class of logic problems, i.e. it is the only operation necessary to solve any problem in the class.

The big head-scratcher with symbolic logic is whether it captures everything about how we communicate. Think about the colors of a sunset or the feeling of a first kiss – they might not fit neatly into symbols. Critics caution that symbolic logic is brilliant but not the only show in town. It should play nice with the other ways we understand conversations and arguments. The roots of symbolic logic stretch way back to thinkers like Aristotle, but it wasn’t until folks like George Boole and Gottlob Frege stepped up in the 1800s that it truly got its wings.

But they require a huge amount of effort by domain experts and software engineers and only work in very narrow use cases. As soon as you generalize the problem, there will be an explosion of new rules to add (remember the cat detection problem?), which will require more human labor. As some AI scientists point out, symbolic AI systems don’t scale. In what follows, we articulate a constitutive account of symbolic reasoning, Perceptual Manipulations Theory, that seeks to elaborate on the cyborg view in exactly this way.

The words sign and symbol derive from Latin and Greek words, respectively, that mean mark or token, as in “take this rose as a token of my esteem.” Both words mean “to stand for something else” or “to represent something else”.

It also empowers applications including visual question answering and bidirectional image-text retrieval. New deep learning approaches based on Transformer models have now eclipsed these earlier symbolic AI approaches and attained state-of-the-art performance in natural language processing. However, Transformer models are opaque and do not yet produce human-interpretable semantic representations for sentences and documents. Instead, they produce task-specific vectors where the meaning of the vector components is opaque.

Understanding that language is a communication function of the human brain clarifies that while training LLMs on language is effective, it oversimplifies the brain’s complexity. To achieve true intelligence in AI, incorporating symbolic reasoning and addressing the need for persistent memory is crucial. By integrating symbolic reasoning into AI, we build on the legacy of brilliant minds like Harold Cohen and push the boundaries of what AI systems can achieve. As we continue researching and developing LLMs, adding symbolic logic middleware represents a significant step forward, enhancing their ability to reason, plan, and understand the world more comprehensively. The advent of the digital computer in the 1940s gave increased attention to the prospects for automated reasoning. Research in artificial intelligence led to the development of efficient algorithms for logical reasoning, highlighted by Robinson’s invention of resolution theorem proving in the 1960s.

Being able to communicate in symbols is one of the main things that make us intelligent. Therefore, symbols have also played a crucial role in the creation of artificial intelligence. We use symbols all the time to define things (cat, car, airplane, etc.) and people (teacher, police, salesperson). Symbols can represent abstract concepts (bank transaction) or things that don’t physically exist (web page, blog post, etc.). Symbols can be organized into hierarchies (a car is made of doors, windows, tires, seats, etc.).

First of all, every deep neural net trained by supervised learning combines deep learning and symbolic manipulation, at least in a rudimentary sense. Because symbolic reasoning encodes knowledge in symbols and strings of characters. In supervised learning, those strings of characters are called labels, the categories by which we classify input data using a statistical model. The output of a classifier (let’s say we’re dealing with an image recognition algorithm that tells us whether we’re looking at a pedestrian, a stop sign, a traffic lane line or a moving semi-truck), can trigger business logic that reacts to each classification. The work in AI started by projects like the General Problem Solver and other rule-based reasoning systems like Logic Theorist became the foundation for almost 40 years of research.

We can think of individual reasoning steps as the atoms out of which proof molecules are built. We say that a set of premises logically entails a conclusion if and only if every world that satisfies the premises also satisfies the conclusion. The AMR is aligned to the terms used in the knowledge graph using entity linking and relation linking modules and is then transformed to a logic representation.5 This logic representation is submitted to the LNN. LNN performs necessary reasoning such as type-based and geographic reasoning to eventually return the answers for the given question. For example, Figure 3 shows the steps of geographic reasoning performed by LNN using manually encoded axioms and DBpedia Knowledge Graph to return an answer. The logic clauses that describe programs are directly interpreted to run the programs specified.

With our NSQA approach , it is possible to design a KBQA system with very little or no end-to-end training data. Currently popular end-to-end trained systems, on the other hand, require thousands of question-answer or question-query pairs – which is unrealistic in most enterprise scenarios. You can foun additiona information about ai customer service and artificial intelligence and NLP. Limitations were discovered in using simple first-order logic to reason about dynamic domains. Problems were discovered both with regards to enumerating the preconditions for an action to succeed and in providing axioms for what did not change after an action was performed. Early work covered both applications of formal reasoning emphasizing first-order logic, along with attempts to handle common-sense reasoning in a less formal manner.

• The ideal, obviously, is to choose assumptions that allow a system to learn flexibly and produce accurate decisions about their inputs.
• LNNs, on the other hand, maintain upper and lower bounds for each variable, allowing the more realistic open-world assumption and a robust way to accommodate incomplete knowledge.
• Ideally, when we have enough sentences, we know exactly how things stand.
• Each sentence divides the set of possible worlds into two subsets, those in which the sentence is true and those in which the sentence is false, as suggested by the following figure.
• Deep learning and neural networks excel at exactly the tasks that symbolic AI struggles with.

Although the prospect of automated reasoning has achieved practical realization only in the last few decades, it is interesting to note that the concept itself is not new. In fact, the idea of building machines capable of logical reasoning has a long tradition. Model checking is the process of examining the set of all worlds to determine logical entailment. To check whether a set of sentences logically entails a conclusion, we use our premises to determine which worlds are possible and then examine those worlds to see whether or not they satisfy our conclusion. If the number of worlds is not too large, this method works well.

Constraint solvers perform a more limited kind of inference than first-order logic. They can simplify sets of spatiotemporal constraints, such as those for RCC or Temporal Algebra, along with solving other kinds of puzzle problems, such as Wordle, Sudoku, cryptarithmetic problems, and so on. Constraint logic programming can be used to solve scheduling problems, for example with constraint handling rules (CHR). Multiple different approaches to represent knowledge and then reason with those representations have been investigated.

Boole gave substance to this dream in the 1800s with the invention of Boolean algebra and with the creation of a machine capable of computing accordingly. Dropping the repeated symbol on the right hand side, we arrive at the conclusion that, if it is Monday and raining, then Mary loves Quincy. In this regard, there is a strong analogy between the methods of Formal Logic and those of high school algebra. To illustrate this analogy, consider the following algebra problem. The form of the argument is the same as in the previous example, but the conclusion is somewhat less believable. The problem in this case is that the use of nothing here is syntactically similar to the use of beer in the preceding example, but in English it means something entirely different.

Similar to the problems in handling dynamic domains, common-sense reasoning is also difficult to capture in formal reasoning. Examples of common-sense reasoning include implicit reasoning about how people think or general knowledge of day-to-day events, objects, and living creatures. This kind of knowledge is taken for granted and not viewed as noteworthy. Although deep learning has historical roots going back decades, neither the term “deep learning” nor the approach was popular just over five years ago, when the field was reignited by papers such as Krizhevsky, Sutskever and Hinton’s now classic (2012) deep network model of Imagenet. So to summarize, one of the main differences between machine learning and traditional symbolic reasoning is how the learning happens.

There are now several efforts to combine neural networks and symbolic AI. One such project is the Neuro-Symbolic Concept Learner (NSCL), a hybrid AI system developed by the MIT-IBM Watson AI Lab. NSCL uses both rule-based programs and neural networks to solve visual question-answering problems. As opposed to pure neural network–based models, the hybrid AI can learn new tasks with less data and is explainable. And unlike symbolic-only models, NSCL doesn’t struggle to analyze the content of images.

Neuro symbolic reasoning and learning is a topic that combines ideas from deep neural networks with symbolic reasoning and learning to overcome several significant technical hurdles such as explainability, modularity, verification, and the enforcement of constraints. While neuro symbolic ideas date back to the early 2000’s, there have been significant advances in the last 5 years. In this chapter, we outline some of these advancements and discuss how they align with several taxonomies for neuro symbolic reasoning. If the capacity for symbolic reasoning is in fact idiosyncratic and context-dependent in the way suggested here, what are the implications for scientific psychology?

We investigate an unconventional direction of research that aims at converting neural networks, a class of distributed, connectionist, sub-symbolic models into a symbolic level with the ultimate goal of achieving AI interpretability and safety. To that end, we propose Object-Oriented Deep Learning, a novel computational paradigm of deep learning that adopts interpretable “objects/symbols” as a basic representational atom instead of N-dimensional tensors (as in traditional “feature-oriented” deep learning). For visual processing, each “object/symbol” can explicitly package common properties of visual objects like its position, pose, scale, probability of being an object, pointers to parts, etc., providing a full spectrum of interpretable visual knowledge throughout all layers.

Moreover, even when we do engage with physical notations, there is a place for semantic metaphors and conscious mathematical rule following. Therefore, although it seems likely that abstract mathematical ability relies heavily on personal histories of active engagement with notational formalisms, this is unlikely to be the story as a whole. It is also why non-human animals, despite in some cases having similar perceptual systems, fail to develop significant mathematical competence even when immersed in a human symbolic environment. Although some animals have been taught to order a small subset of the numerals (less than 10) and carry out simple numerosity tasks within that range, they fail to generalize the patterns required for the indefinite counting that children are capable of mastering, albeit with much time and effort. And without that basis for understanding the domain and range of symbols to which arithmetical operations can be applied, there is no basis for further development of mathematical competence. Perceptual Manipulations Theory claims that symbolic reasoning is implemented over interactions between perceptual and motor processes with real or imagined notational environments.

A different way to create AI was to build machines that have a mind of its own. In contrast, a multi-agent system consists of multiple agents that communicate amongst themselves with some inter-agent communication language such as Knowledge Query and Manipulation Language (KQML). Advantages of multi-agent systems include the ability to divide work among the agents and to increase fault tolerance when agents are lost. Research problems include how agents reach consensus, distributed problem solving, multi-agent learning, multi-agent planning, and distributed constraint optimization. Natural language processing focuses on treating language as data to perform tasks such as identifying topics without necessarily understanding the intended meaning. Natural language understanding, in contrast, constructs a meaning representation and uses that for further processing, such as answering questions.

Most AI approaches make a closed-world assumption that if a statement doesn’t appear in the knowledge base, it is false. LNNs, on the other hand, maintain upper and lower bounds for each variable, allowing the more realistic open-world assumption and a robust way to accommodate incomplete knowledge. At the height of the AI boom, companies such as Symbolics, LMI, and Texas Instruments were selling LISP machines specifically targeted to accelerate the development of AI applications and research. In addition, several artificial intelligence companies, such as Teknowledge and Inference Corporation, were selling expert system shells, training, and consulting to corporations. Symbols also serve to transfer learning in another sense, not from one human to another, but from one situation to another, over the course of a single individual’s life.

Critiques from outside of the field were primarily from philosophers, on intellectual grounds, but also from funding agencies, especially during the two AI winters. The General Problem Solver (GPS) cast planning as problem-solving used means-ends analysis to create plans. STRIPS took a different approach, viewing planning as theorem proving. Graphplan takes a least-commitment approach to planning, rather than sequentially choosing actions from an initial state, working forwards, or a goal state if working backwards.

Samuel’s Checker Program[1952] — Arthur Samuel’s goal was to explore to make a computer learn. The program improved as it played more and Chat GPT more games and ultimately defeated its own creator. In 1959, it defeated the best player, This created a fear of AI dominating AI.

We use it in our professional lives – in proving mathematical theorems, in debugging computer programs, in medical diagnosis, and in legal reasoning. And we use it in our personal lives – in solving puzzles, in playing games, and in doing school assignments, not just in Math but also in History and English and other subjects. But symbolic AI starts to break when you must deal with the messiness of the world. For instance, consider computer vision, the science of enabling computers to make sense of the content of images and video. Say you have a picture of your cat and want to create a program that can detect images that contain your cat. You create a rule-based program that takes new images as inputs, compares the pixels to the original cat image, and responds by saying whether your cat is in those images.

• The form of the argument is the same as in the previous example, but the conclusion is somewhat less believable.
• (1) If a proposition on the left hand side of one sentence is the same as a proposition on the right hand side of the other sentence, it is okay to drop the two symbols, with the proviso that only one such pair may be dropped.
• Each method executes a series of rule-based instructions that might read and change the properties of the current and other objects.
• But in recent years, as neural networks, also known as connectionist AI, gained traction, symbolic AI has fallen by the wayside.

To think that we can simply abandon symbol-manipulation is to suspend disbelief. Similar axioms would be required for other domain actions to specify what did not change. Qualitative simulation, such as Benjamin Kuipers’s QSIM,[90] approximates human reasoning about naive physics, such as what happens when we heat a liquid in a pot on the stove. We expect it to heat and possibly boil over, even though we may not know its temperature, its boiling point, or other details, such as atmospheric pressure. Time periods and titles are drawn from Henry Kautz’s 2020 AAAI Robert S. Engelmore Memorial Lecture[19] and the longer Wikipedia article on the History of AI, with dates and titles differing slightly for increased clarity.

For each of the following sentences, say whether or not it is true in this state of the world. Relational Logic expands upon Propositional Logic by providing a means for explicitly talking about individual objects and their interrelationships (not just monolithic conditions). In order to do so, we expand our language to include object constants and relation constants, variables and quantifiers.