Bibliography

Why I stopped trying to give LLM agents a better memory and instead built Lemmalog, a Datalog engine that maintains an agent's knowledge as analysis state, with provenance, retractions and incremental evaluation, plus what happened when I benchmarked it on LongMemEval and LoCoMo.

The OpenCog (hyper-)graph database and graph rewriting system - opencog/atomspace

Baltag, Moss, and Solecki proposed an expansion of classical modal logic, called logic of epistemic actions and knowledge (EAK), in which one can reason about knowledge and change of knowledge. Kurz and Palmigiano showed how duality theory provides a flexible framework for modeling such epistemic changes, allowing one to develop dynamic epistemic logics on a weaker propositional basis than classical logic (for example an intuitionistic basis). In this paper we show how the techniques of Kurz and Palmigiano can be further extended to define and axiomatize a bilattice logic of epistemic actions and knowledge (BEAK). Our propositional basis is a modal expansion of the well-known four-valued logic of Belnap and Dunn, which is a system designed for handling inconsistent as well as potentially conflicting information. These features, we believe, make our framework particularly promising from a computer science perspective.

This paper concerns static-analysis algorithms for analyzing x86 executables. The aim of the work is to recover intermediate representations that are similar to those that can be created for a program written in a high-level language. Our goal is to perform this task for programs such as plugins, mobile code, worms, and virus-infected code. For such programs, symbol-table and debugging information is either entirely absent, or cannot be relied upon if present; hence, the technique described in the paper makes no use of symbol-table/debugging information. Instead, an analysis is carried out to recover information about the contents of memory locations and how they are manipulated by the executable.

We introduce a framework uniting algorithmic randomness with exchangeable credences to address foundational questions in philosophy of probability and philosophy of science. To demonstrate its power, we show how one might use the framework to derive the Principal Principle -- the norm that rational credence should match known objective chance -- without circularity. The derivation brings together de Finetti's exchangeability, Martin-Löf randomness, Lewis's and Skyrms's chance-credence norms, and statistical constraining laws (arXiv:2303.01411). Laws that constrain histories to algorithmically random sequences naturally pair with exchangeable credences encoding inductive symmetries. Using the de Finetti representation theorem, we show that this pairing directly entails the Principal Principle of this framework. We extend the proof to partial exchangeability and provide finite-history bounds that vanish in the infinite limit. The Principal Principle thus emerges as a mathematical consequence of the alignment between nomological constraints and inductive learning. This reveals how algorithmic randomness and exchangeability can illuminate foundational questions about chance, frequency, and rational belief.

The black swan theory or theory of black swan events is a metaphor that describes an event that comes as a surprise, has a major effect, and is often inappropriately rationalized after the fact with the benefit of hindsight. The term arose from a Latin expression which was based on the presumption that black swans did not exist. The expression was used in the original manner until around 1697 when Dutch mariners saw black swans living in Australia. After this, the term was reinterpreted to mean an unforeseen and consequential event. The reinterpreted theory was articulated by Nassim Nicholas Taleb, starting in 2001, to explain: The disproportionate role of high-profile, hard-to-predict, and rare events that are beyond the realm of normal expectations in history, science, economics, and technology. The non-computability of the probability of consequential rare events using scientific methods (owing to the very nature of small probabilities). The individual and collective psychological biases that make people blind to uncertainty, and to the significant role of rare events in historical affairs. In his 2010 book, Taleb defines the term as an event with two characteristics: first, it is so rare and outside the realm of expectations that it is unpredictable; second, its consequences are extreme—either beneficial or catastrophic—though usually only the catastrophic Black Swan events attract attention. Definitionally, Taleb considers black swans to be in the eye of the beholder and warns that objectively defining a black swan in a way "invariant in the eyes of all observers" would be erroneous. Taleb provides the example of the 9/11 attacks, which were a black swan for many, but not for its planners and perpetrators. Taleb's "black swan theory" (which differs from the earlier philosophical versions of the problem) refers only to statistically unexpected events of large magnitude and consequence and their dominant role in history. Such events, considered extreme outliers, collectively play vastly larger roles than regular occurrences. More technically, in the scientific monograph "Silent Risk", Taleb mathematically defines the black swan problem as "stemming from the use of degenerate metaprobability".

This monograph describes a methodology for the design of knowledge-based systems. A knowledge-based system contains knowledge as well as information and data. The information and data in such a system can be modelled and imple mented as a database. The knowledge in such a system can be implemented either in a programming language or in an expert systems shell. This methodology has two distinguishing features. First, it is "unified". A unified methodology repre sents the data, information and knowledge in a homogeneous manner, as well as the relationships between them. Second, the methodology builds a maintenance mechanism into the design. In knowledge engineering terms, the representation used by this methodology to model knowledge bases applies equally to databases. In database terms, the representation used by this methodology to model databases applies equally to the database rules. The unified methodology unifies the design of the "knowledge base compo nent" and the "database component". "Unification" is achieved in five senses.

In the philosophical literature, the term “abduction” isused in two related but different senses. In both senses, the termrefers to some form of explanatory reasoning. However, in thehistorically first sense, it refers to the place of explanatoryreasoning in generating hypotheses, while in the sense inwhich it is used most frequently in the modern literature it refers tothe place of explanatory reasoning in justifying hypotheses.In the latter sense, abduction is also often called “Inferenceto the Best Explanation.”

There is an ongoing controversy in philosophy about the connection between explanation and inference. According to Bayesians, explanatory considerations should be given weight in determining which inferences to make, if at all, only insofar as doing so is compatible with Strict Conditionalization. Explanationists, on the other hand, hold that explanatory considerations can be relevant to the question of how much confidence to invest in our hypotheses in ways which violate Strict Conditionalization. The controversy has focused on normative issues. This paper investigates experimentally the descriptive question of whether judgments of the explanatory goodness of hypotheses do play a role when people revise their degrees of belief in those hypotheses upon the receipt of new evidence. We present the results of three experiments that together strongly support the predictive superiority of the explanationist position.

Reasoning from inconclusive evidence, or 'induction', is central to science and any applications we make of it. For that reason alone it demands the attention of philosophers of science. This element explores the prospects of using probability theory to provide an inductive logic: a framework for representing evidential support. Constraints on the ideal evaluation of hypotheses suggest that the overall standing of a hypothesis is represented by its probability in light of the total evidence, and incremental support, or confirmation, indicated by the hypothesis having a higher probability conditional on some evidence than it does unconditionally. This proposal is shown to have the capacity to reconstruct many canons of the scientific method and inductive inference. Along the way, significant objections are discussed, such as the challenge of inductive scepticism, and the objection that the probabilistic approach makes evidential support arbitrary.

In artificial intelligence (AI), an expert system is a computer system emulating the decision-making ability of a human expert. Expert systems are designed to solve complex problems by reasoning through bodies of knowledge, represented mainly as if–then rules rather than through conventional procedural programming code. Expert systems were among the first truly successful forms of AI software. They were created in the 1970s and then proliferated in the 1980s, being then widely regarded as the future of AI before the advent of successful artificial neural networks. An expert system is divided into two subsystems: 1) a knowledge base, which represents facts and rules; and 2) an inference engine, which applies the rules to the known facts to deduce new facts, and can include explaining and debugging abilities.

In a variety of emerging applications one needs to decide whether a graph G matches another Gp, i.e., whether G has a topological structure similar to that of Gp. The traditional notions of graph homomorphism and isomorphism often fall short of capturing the structural similarity in these applications. This paper studies revisions of these notions, providing a full treatment from complexity to algorithms. (1) We propose p-homomorphism (p-hom) and 1-1 p-hom, which extend graph homomorphism and subgraph isomorphism, respectively, by mapping edges from one graph to paths in another, and by measuring the similarity of nodes. (2) We introduce metrics to measure graph similarity, and several optimization problems for p-hom and 1-1 p-hom. (3) We show that the decision problems for p-hom and 1-1 p-hom are NP-complete even for DAGs, and that the optimization problems are approximation-hard. (4) Nevertheless, we provide approximation algorithms with provable guarantees on match quality. We experimentally verify the effectiveness of the revised notions and the efficiency of our algorithms in Web site matching, using real-life and synthetic data.

In a variety of emerging applications one needs to decide whether a graph G matches another G p , i.e. , whether G has a topological structure similar to that of G p . The traditional notions of graph homomorphism and isomorphism often fall short of capturing the structural similarity in these applications. This paper studies revisions of these notions, providing a full treatment from complexity to algorithms. (1) We propose p-homomorphism (p -hom) and 1-1 p -hom, which extend graph homomorphism and subgraph isomorphism, respectively, by mapping edges from one graph to paths in another, and by measuring the similarity of nodes . (2) We introduce metrics to measure graph similarity, and several optimization problems for p -hom and 1-1 p -hom. (3) We show that the decision problems for p -hom and 1-1 p -hom are NP-complete even for DAGs, and that the optimization problems are approximation-hard. (4) Nevertheless, we provide approximation algorithms with provable guarantees on match quality. We experimentally verify the effectiveness of the revised notions and the efficiency of our algorithms in Web site matching, using real-life and synthetic data.

The purpose of this monograph is to give an axiomatic foundation for the theory of probability. The author set himself the task of putting in their natural place, among the general notions of modern mathematics, the basic concepts of probability theory—concepts which until recently were considered to be quite peculiar. This task would have been a rather hopeless one before the introduction of Lebesgue’s theories of measure and integration. However, after Lebesgue’s publication of his investigations, the analogies between measure of a set and probability of an event, and between integral of a function and mathematical expectation of a random variable, became apparent. These analogies allowed of further extensions; thus, for example, various properties of independent random variables were seen to be in complete analogy with the corresponding properties of orthogonal functions. But if probability theory was to be based on the above analogies, it still was necessary to make the theories of measure and integra tion independent of the geometric elements which were in the foreground with Lebesgue. This has been done by Frechet. While a conception of probability theory based on the above general viewpoints has been current for some time among certain mathematicians, there was lacking a complete exposition of the whole system, free of extraneous complications. (Cf., however, the book by Frechet, [2] in the bibliography.) I wish to call attention to those points of the present exposition which are outside the above-mentioned range of ideas familiar to the specialist. They are the following: Probability distributions in infinite-dimensional spaces (Chapter III, § 4) ; differentiation and integration of mathematical expectations with respect to a parameter (Chapter IV, § 5) ; and especially the theory of condi tional probabilities and conditional expectations (Chapter V). It should be emphasized that these new problems arose, of neces sity, from some perfectly concrete physical problems. The sixth chapter contains a survey, without proofs, of some results of A. Khinchine and the author of the limitations on the applicability of the ordinary and of the strong law of large num bers. The bibliography contains some recent works which should be of interest from the point of view of the foundations of the subject. I wish to express my warm thanks to Mr. Khinchine, who has read carefully the whole manuscript and proposed several improvements. Kljasma near Moscow, Easter 1933.

To one who has just begun to make his acquaintance with the literature of general semantics, Mother Gorman's book will prove an invaluable guide. From her first chapter giving a historical sketch of the main ideas to her final chapter surveying the ways in which they have influenced education in America, the book is a mine of useful information. Mother Gorman is not a general semanticist. Her reservations about what she regards as the profound philosophical errors of general semantics naturally keep her from aligning herself with this school of thought. But she is an unusually interested bystander and a diligent scholar. Hence she has made an extremely thorough search of the literature, with the result that in many ways she knows a lot more about general semantics than many who call themselves semanticists.--S. I. Hayakawa

In second-order uncertain Bayesian networks, the conditional probabilities are only known within distributions, i.e., probabilities over probabilities. The delta-method has been applied to extend exact first-order inference methods to propagate both means and variances through sum-product networks derived from Bayesian networks, thereby characterizing epistemic uncertainty, or the uncertainty in the model itself. Alternatively, second-order belief propagation has been demonstrated for polytrees but not for general directed acyclic graph structures. In this work, we extend Loopy Belief Propagation to the setting of second-order Bayesian networks, giving rise to Second-Order Loopy Belief Propagation (SOLBP). For second-order Bayesian networks, SOLBP generates inferences consistent with those generated by sum-product networks, while being more computationally efficient and scalable.

Information integration (II) is the merging of information from heterogeneous sources with differing conceptual, contextual and typographical representations. It is used in data mining and consolidation of data from unstructured or semi-structured resources. Typically, information integration refers to textual representations of knowledge but is sometimes applied to rich-media content. Information fusion, which is a related term, involves the combination of information into a new set of information towards reducing redundancy and uncertainty. Examples of technologies available to integrate information include deduplication, and string metrics which allow the detection of similar text in different data sources by fuzzy matching. A host of methods for these research areas are available such as those presented in the International Society of Information Fusion. Other methods rely on causal estimates of the outcomes based on a model of the sources.

Knowledge acquisition is the process used to define the rules and ontologies required for a knowledge-based system. The phrase was first used in conjunction with expert systems to describe the initial tasks associated with developing an expert system, namely finding and interviewing domain experts and capturing their knowledge via rules, objects, and frame-based ontologies. Expert systems were one of the first successful applications of artificial intelligence technology to real world business problems. Researchers at Stanford and other AI laboratories worked with doctors and other highly skilled experts to develop systems that could automate complex tasks such as medical diagnosis. Until this point computers had mostly been used to automate highly data intensive tasks but not for complex reasoning. Technologies such as inference engines allowed developers for the first time to tackle more complex problems. As expert systems scaled up from demonstration prototypes to industrial strength applications it was soon realized that the acquisition of domain expert knowledge was one of if not the most critical task in the knowledge engineering process. This knowledge acquisition process became an intense area of research on its own. One of the earlier works on the topic used Batesonian theories of learning to guide the process. One approach to knowledge acquisition investigated was to use natural language parsing and generation to facilitate knowledge acquisition. Natural language parsing could be performed on manuals and other expert documents and an initial first pass at the rules and objects could be developed automatically. Text generation was also extremely useful in generating explanations for system behavior. This greatly facilitated the development and maintenance of expert systems. A more recent approach to knowledge acquisition is a re-use based approach. Knowledge can be developed in ontologies that conform to standards such as the Web Ontology Language (OWL). In this way knowledge can be standardized and shared across a broad community of knowledge workers. One example domain where this approach has been successful is bioinformatics.

ABSTRACT The article argues for a system of axioms that is meant to capture the logic of normative reasons for belief. The system concerns a primitive direct‐reason relation, a defined doxastic‐reason relation, and a primitive function for the revision of rational belief by reasons. Reasons are assumed to be facts that speak for belief with numerical strength, and the aggregation of non‐overlapping direct reasons involves the intersection of sets of possible worlds and the summing up of strengths. Rational belief is reconstructed as subjective probability, and ratios of new‐to‐old odds are normatively postulated to be a function of the strengths of reason. The resulting theory avoids problems that have been ascribed to the additive aggregation of reasons; it entails that reasons exert epistemic forces that conform to a vector‐like structure and that the rational impact of a reason on rational belief corresponds to a probabilistic Jeffrey/Field‐update. And although the theory differs from existing logics of reason, it is partially continuous with two of them. Its upshot is a systematic bridge between the philosophical theory of epistemic reasons and Bayesian formal epistemology.

This essay develops a joint theory of rational (all-or-nothing) belief and degrees of belief. The theory is based on three assumptions: the logical closure of rational belief; the axioms of probability for rational degrees of belief; and the so-called Lockean thesis, in which the concepts of rational belief and rational degree of belief figure simultaneously. In spite of what is commonly believed, this essay will show that this combination of principles is satisfiable (and indeed nontrivially so) and that the principles are jointly satisfied if and only if rational belief is equivalent to the assignment of a stably high rational degree of belief. Although the logical closure of belief and the Lockean thesis are attractive postulates in themselves, initially this may seem like a formal “curiosity”; however, as will be argued in the rest of the essay, a very reasonable theory of rational belief can be built around these principles that is not ad hoc and that has various philosophical features that are plausible independently. In particular, this essay shows that the theory allows for a solution to the Lottery Paradox, and it has nice applications to formal epistemology. The price that is to be paid for this theory is a strong dependency of belief on the context, where a context involves both the agent's degree of belief function and the partitioning or individuation of the underlying possibilities. But as this essay argues, that price seems to be affordable.This essay develops a joint theory of rational (all-or-nothing) belief and degrees of belief. The theory is based on three assumptions: the logical closure of rational belief; the axioms of probability for rational degrees of belief; and the so-called Lockean thesis, in which the concepts of rational belief and rational degree of belief figure simultaneously. In spite of what is commonly believed, I will show that this combination of principles is satisfiable (and indeed nontrivially so) and that the principles are jointly satisfied if and only if rational belief is equivalent to the assignment of a stably high rational degree of belief. Although the logical closure of belief and the Lockean thesis are attractive postulates in themselves, initially this may seem like a formal “curiosity”; however, as I am going to argue in the rest of the essay, a very reasonable theory of rational belief can be built around these principles that is not ad hoc but that has various philosophical features that are plausible independently.

Currently, it is hard to reap the benefits of deep learning for Bayesian methods, which allow the explicit specification of prior knowledge and accurately capture model uncertainty. We present Prior-Data Fitted Networks (PFNs). PFNs leverage in-context learning in large-scale machine learning techniques to approximate a large set of posteriors. The only requirement for PFNs to work is the ability to sample from a prior distribution over supervised learning tasks (or functions). Our method restates the objective of posterior approximation as a supervised classification problem with a set-valued input: it repeatedly draws a task (or function) from the prior, draws a set of data points and their labels from it, masks one of the labels and learns to make probabilistic predictions for it based on the set-valued input of the rest of the data points. Presented with a set of samples from a new supervised learning task as input, PFNs make probabilistic predictions for arbitrary other data points in a single forward propagation, having learned to approximate Bayesian inference. We demonstrate that PFNs can near-perfectly mimic Gaussian processes and also enable efficient Bayesian inference for intractable problems, with over 200-fold speedups in multiple setups compared to current methods. We obtain strong results in very diverse areas such as Gaussian process regression, Bayesian neural networks, classification for small tabular data sets, and few-shot image classification, demonstrating the generality of PFNs. Code and trained PFNs are released at https://github.com/automl/TransformersCanDoBayesianInference.

The OpenCog (hyper-)graph database and graph rewriting system

The OpenCog (hyper-)graph database and graph rewriting system - opencog/atomspace

Recent iPhone models have additional hardware-based security protection for sensitive regions of the kernel memory. We discovered that to bypass this hardware-based security protection, the attackers used another hardware feature of Apple-designed SoCs.

PDF | The article argues for a system of axioms that is meant to capture the logic of normative reasons for belief. The system concerns a primitive... | Find, read and cite all the research you need on ResearchGate

Accurate and robust disassembly of stripped binaries is challenging. The root of the difficulty is that high-level structures, such as instruction and function boundaries, are absent in stripped binaries and must be recovered based on incomplete information. Current disassembly approaches rely on heuristics or simple pattern matching to approximate the recovery, but these methods are often inaccurate and brittle, especially across different compiler optimizations. We present XDA, a transfer-learning-based disassembly framework that learns different contextual dependencies present in machine code and transfers this knowledge for accurate and robust disassembly. We design a self-supervised learning task motivated by masked Language Modeling to learn interactions among byte sequences in binaries. The outputs from this task are byte embeddings that encode sophisticated contextual dependencies between input binaries' byte tokens, which can then be finetuned for downstream disassembly tasks. We evaluate XDA's performance on two disassembly tasks, recovering function boundaries and assembly instructions, on a collection of 3,121 binaries taken from SPEC CPU2017, SPEC CPU2006, and the BAP corpus. The binaries are compiled by GCC, ICC, and MSVC on x86/x64 Windows and Linux platforms over 4 optimization levels. XDA achieves 99.0% and 99.7% F1 score at recovering function boundaries and instructions, respectively, surpassing the previous state-of-the-art on both tasks. It also maintains speed on par with the fastest ML-based approach and is up to 38x faster than hand-written disassemblers like IDA Pro. We release the code of XDA at https://github.com/CUMLSec/XDA.

A priori (‘from the earlier’) and a posteriori (‘from the later’) are Latin phrases used in philosophy and linguistics to distinguish types of knowledge, justification, or argument by their reliance on experience. Roughly speaking, a proposition is known or justified a priori if it is known or justified independently of any experience (beyond the experience necessary to understand the proposition); instead, it is known or justified a posteriori if its knowledge and/or justification depends on empirical evidence. For example, the proposition ‘It is sunny in London today’ can be known (if true) a posteriori, whereas the proposition ‘Either it is sunny or it is not sunny in London today’ can be known a priori. Fields of knowledge where a priori justification is predominant are, for example, mathematics and formal logic; by contrast, most of the sciences generally involve a posteriori justification. In the history of philosophy, the a priori–a posteriori distinction first appeared in the writings of the 14th century logician Albert of Saxony, where the phrases were used to distinguish between arguments ‘from causes to effects’ (a priori) and ‘from effects to causes’ (a posteriori). As an epistemological distinction it became prominent with Immanuel Kant's Critique of Pure Reason, where its relation to the analytic–synthetic distinction is discussed.

The standard probability axioms are the foundations of probability theory introduced by Russian mathematician Andrey Kolmogorov in 1933. Like all axiomatic systems, they outline the basic assumptions underlying the application of probability to fields such as pure mathematics and the physical sciences, while avoiding logical paradoxes. The probability axioms do not specify or assume any particular interpretation of probability, but may be motivated by starting from a philosophical definition of probability and arguing that the axioms are satisfied by this definition. For example, Cox's theorem derives the laws of probability based on a "logical" definition of probability as the likelihood or credibility of arbitrary logical propositions. The Dutch book arguments show that rational agents must make bets which are in proportion with a subjective measure of the probability of events. The third axiom, σ-additivity, is relatively modern, and originates with Lebesgue's measure theory. Some authors replace this with the strictly weaker axiom of finite additivity, which is sufficient to deal with some applications.

In propositional calculus, a propositional function or a predicate is a sentence expressed in a way that would assume the value of true or false, except that within the sentence there is a variable (x) that is not defined or specified (thus being a free variable), which leaves the statement undetermined. The sentence may contain several such variables (e.g. n variables, in which case the function takes n arguments).

Epistemic logic is a subfield of philosophical logic concerned withlogical approaches to knowledge, belief, and related notions. Thoughany logic with an epistemic interpretation may be called anepistemic logic, the most widespread type of epistemic logicsin use at present are modal logics. Knowledge and belief arerepresented via the modal operators K and B, often witha subscript indicating the agent that holds the attitude. FormulasKaφKaφK_{a}\varphi and BaφBaφB_{a}\varphi are then read “agenta knows that phi” and “agent a believes thatphi”, respectively. Epistemic logic allows the formalexploration of the implications of epistemic principles. For example,the formula Kaφ→φKaφ→φK_{a}\varphi\rightarrow\varphi states that what isknown is true, while Kaφ→KaKaφKaφ→KaKaφK_{a}\varphi\rightarrow K_{a}K_{a}\varphistates that what is known is known to be known. The semantics ofepistemic logic are typically given in terms of possible worldsvia Kripke models such that the formula KaφKaφK_{a}\varphi isread to assert that φφ\varphi is true in all worlds agent aconsiders epistemically possible relative to its current information.The central problems that have concerned epistemic logicians include,for example, determining which epistemic principles are mostappropriate for characterizing knowledge and belief, the logicalrelations between different conceptions of knowledge and belief, andthe epistemic features of groups of agents. Beyond philosophy proper,epistemic logic flourishes in theoretical computer science, AI,economics, and related fields.

A semantic reasoner, reasoning engine, rules engine, or simply a reasoner, is a piece of software able to infer logical consequences from a set of asserted facts or axioms. The notion of a semantic reasoner generalizes that of an inference engine, by providing a richer set of mechanisms to work with. The inference rules are commonly specified by means of an ontology language, and often a description logic language. Many reasoners use first-order predicate logic to perform reasoning; inference commonly proceeds by forward chaining and backward chaining. There are also examples of probabilistic reasoners, including non-axiomatic reasoning systems, and probabilistic logic networks.

Predicate logic based reasoning approaches provide a means of formally specifying domain knowledge and manipulating symbolic information to explicitly reason about different concepts of interest. Extension of traditional binary predicate logics with the bilattice formalism permits the handling of uncertainty in reasoning, thereby facilitating their application to computer vision problems. In this paper, we propose using first order predicate logics, extended with a bilattice based uncertainty handling formalism, as a means of formally encoding pattern grammars, to parse a set of image features, and detect the presence of different patterns of interest. Detections from low level feature detectors are treated as logical facts and, in conjunction with logical rules, used to drive the reasoning. Positive and negative information from different sources, as well as uncertainties from detections, are integrated within the bilattice framework. We show that this approach can also generate proofs or justifications (in the form of parse trees) for each hypothesis it proposes thus permitting direct analysis of the final solution in linguistic form. Automated logical rule weight learning is an important aspect of the application of such systems in the computer vision domain. We propose a rule weight optimization method which casts the instantiated inference tree as a knowledge-based neural network, interprets rule uncertainties as link weights in the network, and applies a constrained, back-propagation algorithm to converge upon a set of rule weights that give optimal performance within the bilattice framework. Finally, we evaluate the proposed predicate logic based pattern grammar formulation via application to the problems of (a) detecting the presence of humans under partial occlusions and (b) detecting large complex man made structures as viewed in satellite imagery. We also evaluate the optimization approach on real as well as simulated data and show favorable results.

'Fundamentals of Bayesian Epistemology' provides an accessible introduction to the key concepts and principles of the Bayesian formalism. This volume introduces degrees of belief as a concept in epistemology and the rules for updating degrees of belief derived from Bayesian principles

The Welch Labs Illustrated Guide to AI offers unique perspectives on artificial intelligence through hands-on exploration and richly detailed graphics that give readers a visceral understanding of how AI actually works. From the fundamental Perceptron to cutting-edge AI video generation, each chapter combines thought-provoking exercises with supporting code to illuminate the key breakthroughs and persistent mysteries in AI development. The book is a great fit for both seasoned professionals and students.

A knowledge graph represents a network of real-world entities—such as objects, events, situations or concepts—and illustrates the relationship between them.

The raven paradox, also known as Hempel's paradox, Hempel's ravens or, rarely, the paradox of indoor ornithology, is a paradox arising from the question of what constitutes evidence for the truth of a statement. Observing objects that are neither black nor ravens may formally increase the likelihood that all ravens are black even though, intuitively, these observations are unrelated. This problem was proposed by the logician Carl Gustav Hempel in the 1940s to illustrate a contradiction between inductive logic and intuition.

Binary analysis, a cornerstone technique in cybersecurity, enables the examination of binary executables, irrespective of source code availability. It plays a critical role in understanding program behaviors, detecting software bugs, and mitigating potential vulnerabilities, specially in situations where the source code remains out of reach. However, aligning the efficacy of binary analysis with that of source-level analysis remains a significant challenge, primarily due to the uncertainty caused by the loss of semantic information during the compilation process. This dissertation presents an innovative probabilistic approach, termed as probabilistic binary analysis, designed to combat the intrinsic uncertainty in binary analysis. It builds on the fundamental principles of program sampling and probabilistic inference, enhanced further by an iterative refinement architecture. The dissertation suggests that a thorough and practical method of sampling program behaviors can yield a substantial quantity of hints which could be instrumental in recovering lost information, despite the potential inclusion of some inaccuracies. Consequently, a probabilistic inference technique is applied to systematically incorporate and process the collected hints, suppressing the incorrect ones, thereby enabling the interpretation of high-level semantics. Furthermore, an iterative refinement mechanism is deployed to augment the efficiency of the probabilistic analysis in subsequent applications, facilitating the progressive enhancement of analysis outcomes through an automated or human-guided feedback loop. This work offers an in-depth understanding of the challenges and solutions related to assessing low-level program representations and systematically handling the inherent uncertainty in binary analysis. It aims to contribute to the field by advancing the development of precise, reliable, and interpretable binary analysis solutions, thereby setting the groundwork for future exploration in this domain.

A Datalog engine for LLM agent memory: stratified rules, provenance-tracked facts, incremental derivation, and an MCP server that lets your harness use it as a shared brain.