Reasoning, a crucial ability for complex problem-solving, plays a pivotal role in various real-world settings such as negotiation, medical diagnosis, and criminal investigation. It serves as a fundamental methodology in the field of Artificial General Intelligence (AGI). With the ongoing development of foundation models, there is a growing interest in exploring their abilities in reasoning tasks. In this paper, we introduce seminal foundation models proposed or adaptable for reasoning, highlighting the latest advancements in various reasoning tasks, methods, and benchmarks. We then delve into the potential future directions behind the emergence of reasoning abilities within foundation models. We also discuss the relevance of multimodal learning, autonomous agents, and super alignment in the context of reasoning. By discussing these future research directions, we hope to inspire researchers in their exploration of this field, stimulate further advancements in reasoning with foundation models, and contribute to the development of AGI.
As a rule statistical measures are often vulnerable to the presence of outliers and spatial correlation coefficients, critical in the assessment of spatial data, remain susceptible to this inherent flaw. In contexts where data originates from a variety of domains (such as, e. g., socio-economic, environmental or epidemiological disciplines) it is quite common to encounter not just anomalous data points, but also non-normal distributions. These irregularities can significantly distort the broader analytical landscape often masking significant spatial attributes. This paper embarks on a mission to enhance the resilience of traditional spatial correlation metrics, specifically the Moran Coefficient (MC), Geary's Contiguity ratio (GC), and the Approximate Profile Likelihood Estimator (APLE) and to propose a series of alternative measures. Drawing inspiration from established analytical paradigms, our research harnesses the power of influence function studies to examine the robustness of the proposed novel measures in the presence of different outlier scenarios.
Honeypots are essential tools in cybersecurity. However, most of them (even the high-interaction ones) lack the required realism to engage and fool human attackers. This limitation makes them easily discernible, hindering their effectiveness. This work introduces a novel method to create dynamic and realistic software honeypots based on Large Language Models. Preliminary results indicate that LLMs can create credible and dynamic honeypots capable of addressing important limitations of previous honeypots, such as deterministic responses, lack of adaptability, etc. We evaluated the realism of each command by conducting an experiment with human attackers who needed to say if the answer from the honeypot was fake or not. Our proposed honeypot, called shelLM, reached an accuracy of 0.92. The source code and prompts necessary for replicating the experiments have been made publicly available.
The sumcheck protocol, introduced in 1992, is an interactive proof which is a key component of many probabilistic proof systems in computational complexity theory and cryptography, some of which have been deployed. However, none of these proof systems based on the sumcheck protocol enjoy a formally-verified security analysis. In this paper, we make progress in this direction by providing a formally verified security analysis of the sumcheck protocol using the interactive theorem prover Isabelle/HOL. We follow a general and modular approach. First, we give a general formalization of public-coin interactive proofs. We then define a generalized sumcheck protocol for which we axiomatize the underlying mathematical structure and we establish its soundness and completeness. Finally, we prove that these axioms hold for multivariate polynomials, the original setting of the sumcheck protocol. Our modular analysis facilitates formal verification of sumcheck instances based on different mathematical structures with little effort, by simply proving that these structures satisfy the axioms. Moreover, the analysis supports the development and formal verification of future cryptographic protocols using the sumcheck protocol as a building block.
In many classification applications, the prediction of a deep neural network (DNN) based classifier needs to be accompanied with some confidence indication. Two popular post-processing approaches for that aim are: 1) calibration: modifying the classifier's softmax values such that their maximum (associated with the prediction) better estimates the correctness probability; and 2) conformal prediction (CP): devising a score (based on the softmax values) from which a set of predictions with theoretically guaranteed marginal coverage of the correct class is produced. While in practice both types of indications can be desired, so far the interplay between them has not been investigated. Toward filling this gap, in this paper we study the effect of temperature scaling, arguably the most common calibration technique, on prominent CP methods. We start with an extensive empirical study that among other insights shows that, surprisingly, calibration has a detrimental effect on popular adaptive CP methods: it frequently leads to larger prediction sets. Then, we turn to theoretically analyze this behavior. We reveal several mathematical properties of the procedure, according to which we provide a reasoning for the phenomenon. Our study suggests that it may be worthwhile to utilize adaptive CP methods, chosen for their enhanced conditional coverage, based on softmax values prior to (or after canceling) temperature scaling calibration.
Natural policy gradient (NPG) methods with entropy regularization achieve impressive empirical success in reinforcement learning problems with large state-action spaces. However, their convergence properties and the impact of entropy regularization remain elusive in the function approximation regime. In this paper, we establish finite-time convergence analyses of entropy-regularized NPG with linear function approximation under softmax parameterization. In particular, we prove that entropy-regularized NPG with averaging satisfies the \emph{persistence of excitation} condition, and achieves a fast convergence rate of $\tilde{O}(1/T)$ up to a function approximation error in regularized Markov decision processes. This convergence result does not require any a priori assumptions on the policies. Furthermore, under mild regularity conditions on the concentrability coefficient and basis vectors, we prove that entropy-regularized NPG exhibits \emph{linear convergence} up to a function approximation error.
Spoken dialogue systems have transformed human-machine interaction by providing real-time responses to queries. However, misunderstandings between the user and system persist. This study explores the significance of interactional language in dialogue repair between virtual assistants and users by analyzing interactions with Google Assistant and Siri, focusing on their utilization and response to the other-initiated repair strategy "huh?" prevalent in human-human interaction. Findings reveal several assistant-generated strategies but an inability to replicate human-like repair strategies such as "huh?". English and Spanish user acceptability surveys show differences in users' repair strategy preferences and assistant usage, with both similarities and disparities among the two surveyed languages. These results shed light on inequalities between interactional language in human-human interaction and human-machine interaction, underscoring the need for further research on the impact of interactional language in human-machine interaction in English and beyond.
Driven by the rapid ascent of artificial intelligence (AI), organizations are at the epicenter of a seismic shift, facing a crucial question: How can AI be successfully integrated into existing operations? To help answer it, manage expectations and mitigate frustration, this article introduces Computational Management, a systematic approach to task automation for enhancing the ability of organizations to harness AI's potential within existing workflows. Computational Management acts as a bridge between the strategic insights of management science with the analytical rigor of computational thinking. The article offers three easy step-by-step procedures to begin the process of implementing AI within a workflow. Such procedures focus on task (re)formulation, on the assessment of the automation potential of tasks, on the completion of task specification templates for AI selection and adaptation. Included in the article there are manual and automated methods, with prompt suggestions for publicly available LLMs, to complete these three procedures. The first procedure, task (re)formulation, focuses on breaking down work activities into basic units, so they can be completed by one agent, involve a single well-defined action, and produce a distinct outcome. The second, allows the assessment of the granular task and its suitability for automation, using the Task Automation Index to rank tasks based on whether they have standardized input, well-defined rules, repetitiveness, data dependency, and objective outputs. The third, focuses on a task specification template which details information on 16 critical components of tasks, and can be used as a checklist to select or adapt the most suitable AI solution for integration into existing workflows. Computational Management provides a roadmap and a toolkit for humans and AI to thrive together, while enhancing organizational efficiency and innovation.
The concept of causality plays an important role in human cognition . In the past few decades, causal inference has been well developed in many fields, such as computer science, medicine, economics, and education. With the advancement of deep learning techniques, it has been increasingly used in causal inference against counterfactual data. Typically, deep causal models map the characteristics of covariates to a representation space and then design various objective optimization functions to estimate counterfactual data unbiasedly based on the different optimization methods. This paper focuses on the survey of the deep causal models, and its core contributions are as follows: 1) we provide relevant metrics under multiple treatments and continuous-dose treatment; 2) we incorporate a comprehensive overview of deep causal models from both temporal development and method classification perspectives; 3) we assist a detailed and comprehensive classification and analysis of relevant datasets and source code.
Residual networks (ResNets) have displayed impressive results in pattern recognition and, recently, have garnered considerable theoretical interest due to a perceived link with neural ordinary differential equations (neural ODEs). This link relies on the convergence of network weights to a smooth function as the number of layers increases. We investigate the properties of weights trained by stochastic gradient descent and their scaling with network depth through detailed numerical experiments. We observe the existence of scaling regimes markedly different from those assumed in neural ODE literature. Depending on certain features of the network architecture, such as the smoothness of the activation function, one may obtain an alternative ODE limit, a stochastic differential equation or neither of these. These findings cast doubts on the validity of the neural ODE model as an adequate asymptotic description of deep ResNets and point to an alternative class of differential equations as a better description of the deep network limit.
While it is nearly effortless for humans to quickly assess the perceptual similarity between two images, the underlying processes are thought to be quite complex. Despite this, the most widely used perceptual metrics today, such as PSNR and SSIM, are simple, shallow functions, and fail to account for many nuances of human perception. Recently, the deep learning community has found that features of the VGG network trained on the ImageNet classification task has been remarkably useful as a training loss for image synthesis. But how perceptual are these so-called "perceptual losses"? What elements are critical for their success? To answer these questions, we introduce a new Full Reference Image Quality Assessment (FR-IQA) dataset of perceptual human judgments, orders of magnitude larger than previous datasets. We systematically evaluate deep features across different architectures and tasks and compare them with classic metrics. We find that deep features outperform all previous metrics by huge margins. More surprisingly, this result is not restricted to ImageNet-trained VGG features, but holds across different deep architectures and levels of supervision (supervised, self-supervised, or even unsupervised). Our results suggest that perceptual similarity is an emergent property shared across deep visual representations.