AI Problems Research Articles

Collaborative Intelligence for Safety-Critical Industries: A Literature Review

While AI-driven automation can increase the performance and safety of systems, humans should not be replaced in safety-critical systems but should be integrated to collaborate and mitigate each other’s limitations. The current trend in Industry 5.0 is towards human-centric collaborative paradigms, with an emphasis on collaborative intelligence (CI) or Hybrid Intelligent Systems. In this survey, we search and review recent work that employs AI methods for collaborative intelligence applications, specifically those that focus on safety and safety-critical industries. We aim to contribute to the research landscape and industry by compiling and analyzing a range of scenarios where AI can be used to achieve more efficient human–machine interactions, improved collaboration, coordination, and safety. We define a domain-focused taxonomy to categorize the diverse CI solutions, based on the type of collaborative interaction between intelligent systems and humans, the AI paradigm used and the domain of the AI problem, while highlighting safety issues. We investigate 91 articles on CI research published between 2014 and 2023, providing insights into the trends, gaps, and techniques used, to guide recommendations for future research opportunities in the fast developing collaborative intelligence field.

Enhancing deep reinforcement learning for scale flexibility in real-time strategy games

Real-time strategy (RTS) games present a unique challenge for AI agents due to the combination of several fundamental AI problems. While Deep Reinforcement Learning (DRL) has shown promise in the development of autonomous agents for the genre, existing architectures often struggle with games featuring maps of varying dimensions. This limitation hinders the agent’s ability to generalize its learned strategies across different scenarios. This paper proposes a novel approach that overcomes this problem by incorporating Spatial Pyramid Pooling (SPP) within a DRL framework. We leverage the GridNet architecture’s encoder–decoder structure and integrate an SPP layer into the critic network of the Proximal Policy Optimization (PPO) algorithm. This SPP layer dynamically generates a standardized representation of the game state, regardless of the initial observation size. This allows the agent to effectively adapt its decision-making process to any map configuration. Our evaluations demonstrate that the proposed method significantly enhances the model’s flexibility and efficiency in training agents for various RTS game scenarios, albeit with some discernible limitations when applied to very small maps. This approach paves the way for more robust and adaptable AI agents capable of excelling in sequential decision problems with variable-size observations.

Towards Unlocking the Hidden Potentials of the Data-Centric AI Paradigm in the Modern Era

Data-centric artificial intelligence (DC-AI) is a modern paradigm that gives more priority to data quality enhancement, rather than only optimizing the complex codes of AI models. The DC-AI paradigm is expected to substantially advance the status of AI research and developments, which has been solely based on model-centric AI (MC-AI) over the past 30 years. Until present, there exists very little knowledge about DC-AI, and its significance in terms of solving real-world problems remains unexplored in the recent literature. In this technical note, we present the core aspects of DC-AI and MC-AI and discuss their interplay when used to solve some real-world problems. We discuss the potential scenarios/situations that require the integration of DC-AI with MC-AI to solve challenging problems in AI. We performed a case study on a real-world dataset to corroborate the potential of DC-AI in realistic scenarios and to prove its significance over MC-AI when either data are limited or their quality is poor. Afterward, we comprehensively discuss the challenges that currently hinder the realization of DC-AI, and we list promising avenues for future research and development concerning DC-AI. Lastly, we discuss the next-generation computing for DC-AI that can foster DC-AI-related developments and can help transition DC-AI from theory to practice. Our detailed analysis can guide AI practitioners toward exploring the undisclosed potential of DC-AI in the current AI-driven era.

Expropriated Minds: On Some Practical Problems of Generative AI, Beyond Our Cognitive Illusions

This paper discusses some societal implications of the most recent and publicly discussed application of advanced machine learning techniques: generative AI models, such as ChatGPT (text generation) and DALL-E (text-to-image generation). The aim is to shift attention away from conceptual disputes, e.g. regarding their level of intelligence and similarities/differences with human performance, to focus instead on practical problems, pertaining the impact that these technologies might have (and already have) on human societies. After a preliminary clarification of how generative AI works (Sect. 1), the paper discusses what kind of transparency ought to be required for such technologies and for the business model behind their commercial exploitation (Sect. 2), what is the role of user-generated data in determining their performance and how it should inform the redistribution of the resulting benefits (Sect. 3), the best way of integrating generative AI systems in the creative job market and how to properly negotiate their role in it (Sect. 4), and what kind of “cognitive extension” offered by these technologies we ought to embrace, and what type we should instead resist and monitor (Sect. 5). The last part of the paper summarizes the main conclusions of this analysis, also marking its distance from other, more apocalyptic approaches to the dangers of AI for human society.

A noise audit of human-labeled benchmarks for machine commonsense reasoning

With the advent of large language models, evaluating and benchmarking these systems on important AI problems has taken on newfound importance. Such benchmarking typically involves comparing the predictions of a system against human labels (or a single ‘ground-truth’). However, much recent work in psychology has suggested that most tasks involving significant human judgment can have non-trivial degrees of noise. In his book, Kahneman suggests that noise may be a much more significant component of inaccuracy compared to bias, which has been studied more extensively in the AI community. This article proposes a detailed noise audit of human-labeled benchmarks in machine commonsense reasoning, an important current area of AI research. We conduct noise audits under two important experimental conditions: one in a smaller-scale but higher-quality labeling setting, and another in a larger-scale, more realistic online crowdsourced setting. Using Kahneman’s framework of noise, our results consistently show non-trivial amounts of level, pattern, and system noise, even in the higher-quality setting, with comparable results in the crowdsourced setting. We find that noise can significantly influence the performance estimates that we obtain of commonsense reasoning systems, even if the ‘system’ is a human; in some cases, by almost 10 percent. Labeling noise also affects performance estimates of systems like ChatGPT by more than 4 percent. Our results suggest that the default practice in the AI community of assuming and using a ‘single’ ground-truth, even on problems requiring seemingly straightforward human judgment, may warrant empirical and methodological re-visiting.

Optimization and Performance Evaluation of Artificial Intelligence Algorithms in Game Design

This article aims to optimize artificial intelligence algorithms in game design to enhance game performance and player experience. Therefore, an innovative method called Dynamic Game AI Fitness Optimization Algorithm (DGAFOA) was proposed in the article, and its effectiveness was verified through experiments. In the experimental section, this article applies the DGAFOA algorithm to a typical role-playing game and compares it with traditional fixed parameter AI algorithms. The experimental results show that the DGAFOA algorithm exhibits significant advantages in key indicators such as game task completion rate and player satisfaction. Specifically, game AI using the DGAFOA algorithm can respond more quickly and accurately to player behavior, improving the overall smoothness and fun of the game. In addition, the DGAFOA algorithm also introduces ε- The greedy strategy balances exploration and utilization, effectively avoiding the problem of AI getting stuck in local optima. This enables game AI to maintain stable performance while still possessing the ability to explore new strategies, bringing players more surprises and challenges.

Comparing Student and Writing Instructor Perceptions of Academic Dishonesty When Collaborators Are Artificial Intelligence or Human

It remains unclear if perceptions of academic dishonesty concerning artificial intelligence writing technologies (AIWTs) present new challenges or if they reflect prior, non-AI concerns. To structure this problem, we used a randomized control survey experiment. We compared student ( n = 603) and instructor ( n = 312) attitudes toward dishonesty in collaborations involving humans versus AIWT in 10 writing-related scenarios. Results suggest similar perception patterns among students and instructors, with both populations expressing significant differences in perceived dishonesty between AI and human collaborators in some scenarios. This experiment structures the problem of AI writing and academic dishonesty for future research in this emerging field.

Enhancing secure financial transactions through the synergy of blockchain and artificial intelligence

Blockchain and artificial intelligence are innovative technologies that can securely process and share data across unreliable networks. Due to data leakage from user information, it is critical to keep the data confidential and completely protected because criminals are looking for this information to attack the system or steal information in the banking sector. To address this issue, this paper suggests an Integrated Blockchain and Artificial intelligence (IBAI) Framework for secure financial transactions. A blockchain can store every customer's data in one place, while AI-driven algorithms can speedily examine that data and make an unbiased decision. The rising blockchain technology provides a decentralized architecture that enables the secure sharing of data and resources to the different networks and is promoted for removing centralized control and resolving the problems of AI. When suspicious behaviour happens, alerts can be triggered to avert theft. Furthermore, registration protocols based on AI are utilized to keep this data in comprehensive security. The numerical results show that the suggested IBAI model enhances the suspicious behaviour detection ratio and increases accuracy up to 98% compared to other models.

God from the machine 2.0: biblical narrative as a metalanguage for discussing machine ethics

The article is devoted to theoretical models in which the biblical narrative becomes a means of conceptualizing the actions of technological objects. Based on the works of Russian and foreign philosophers, it is shown what the specifics of the discursive formulation of the problems of machine ethics, understood as a subject, are. The focus is not on the formulation of the problem of artificial intelligence ethics per se or the possibility of solving it, but on the very method of discursive formulation, or the way of talking about it, by incorporating biblical plots, motifs, and images as mechanisms for understanding this problem. The subject of the study is the heuristic potential of the biblical metanarrative in the structure of statements about artificial intelligence. It is demonstrated that the heuristic potential of the biblical metanarrative in the study of AI problems is related to the production of distinctions. It is revealed which new systems of basic distinctions can be constructed using this method of discursive design. It is shown what changes are observed in the discourse about algorithmic technologies. Possible reasons for the activation of biblical discourse in addressing the problems of artificial intelligence are identified.

Solving Satisfiability Modulo Counting for Symbolic and Statistical AI Integration with Provable Guarantees

Satisfiability Modulo Counting (SMC) encompasses problems that require both symbolic decision-making and statistical reasoning. Its general formulation captures many real-world problems at the intersection of symbolic and statistical AI. SMC searches for policy interventions to control probabilistic outcomes. Solving SMC is challenging because of its highly intractable nature (NP^PP-complete), incorporating statistical inference and symbolic reasoning. Previous research on SMC solving lacks provable guarantees and/or suffers from suboptimal empirical performance, especially when combinatorial constraints are present. We propose XOR-SMC, a polynomial algorithm with access to NP-oracles, to solve highly intractable SMC problems with constant approximation guarantees. XOR-SMC transforms the highly intractable SMC into satisfiability problems by replacing the model counting in SMC with SAT formulae subject to randomized XOR constraints. Experiments on solving important SMC problems in AI for social good demonstrate that XOR-SMC outperforms several baselines both in solution quality and running time.

Counterfactual Explanations for Misclassified Images: How Human and Machine Explanations Differ (Abstract Reprint)

Counterfactual explanations have emerged as a popular solution for the eXplainable AI (XAI) problem of elucidating the predictions of black-box deep-learning systems because people easily understand them, they apply across different problem domains and seem to be legally compliant. Although over 100 counterfactual methods exist in the XAI literature, each claiming to generate plausible explanations akin to those preferred by people, few of these methods have actually been tested on users (∼7%). Even fewer studies adopt a user-centered perspective; for instance, asking people for their counterfactual explanations to determine their perspective on a “good explanation”. This gap in the literature is addressed here using a novel methodology that (i) gathers human-generated counterfactual explanations for misclassified images, in two user studies and, then, (ii) compares these human-generated explanations to computationally-generated explanations for the same misclassifications. Results indicate that humans do not “minimally edit” images when generating counterfactual explanations. Instead, they make larger, “meaningful” edits that better approximate prototypes in the counterfactual class. An analysis based on “explanation goals” is proposed to account for this divergence between human and machine explanations. The implications of these proposals for future work are discussed.

MobileInst: Video Instance Segmentation on the Mobile

Video instance segmentation on mobile devices is an important yet very challenging edge AI problem. It mainly suffers from (1) heavy computation and memory costs for frame-by-frame pixel-level instance perception and (2) complicated heuristics for tracking objects. To address these issues, we present MobileInst, a lightweight and mobile-friendly framework for video instance segmentation on mobile devices. Firstly, MobileInst adopts a mobile vision transformer to extract multi-level semantic features and presents an efficient query-based dual-transformer instance decoder for mask kernels and a semantic-enhanced mask decoder to generate instance segmentation per frame. Secondly, MobileInst exploits simple yet effective kernel reuse and kernel association to track objects for video instance segmentation. Further, we propose temporal query passing to enhance the tracking ability for kernels. We conduct experiments on COCO and YouTube-VIS datasets to demonstrate the superiority of MobileInst and evaluate the inference latency on one single CPU core of the Snapdragon 778G Mobile Platform, without other methods of acceleration. On the COCO dataset, MobileInst achieves 31.2 mask AP and 433 ms on the mobile CPU, which reduces the latency by 50% compared to the previous SOTA. For video instance segmentation, MobileInst achieves 35.0 AP and 30.1 AP on YouTube-VIS 2019 & 2021.

Inference and Learning in Dynamic Decision Networks Using Knowledge Compilation

Decision making under uncertainty in dynamic environments is a fundamental AI problem in which agents need to determine which decisions (or actions) to make at each time step to maximise their expected utility. Dynamic decision networks (DDNs) are an extension of dynamic Bayesian networks with decisions and utilities. DDNs can be used to compactly represent Markov decision processes (MDPs). We propose a novel algorithm called mapl-cirup that leverages knowledge compilation techniques developed for (dynamic) Bayesian networks to perform inference and gradient-based learning in DDNs. Specifically, we knowledge-compile the Bellman update present in DDNs into dynamic decision circuits and evaluate them within an (algebraic) model counting framework. In contrast to other exact symbolic MDP approaches, we obtain differentiable circuits that enable gradient-based parameter learning.

VCSEL-based photonic spiking neural networks for ultrafast detection and tracking

Inspired by efficient biological spike-based neural networks, we demonstrate for the first time the detection and tracking of target patterns in image and video inputs at high-speed rates with networks of multiple artificial spiking optical neurons. Using photonic systems of in-parallel spiking vertical cavity surface emitting lasers (VCSELs), we demonstrate the implementation of multiple convolutional kernel operators which, in combination with optical spike signalling, enable the detection and tracking of target features in images/video feeds at an ultrafast photonic operation speed of 1 ns per pixel. Alongside a single layer optical spiking neural network (SNN) demonstration, a multi-layer network of photonic (GHz-rate) spike-firing neurons is reported where the photonic system successfully tracks a large complex feature (Handwritten Digit 3). The consecutive photonic layers perform spike-enabled image reduction and convolution operations, and interact with a software-implemented SNN, that learns the feature patterns that best identify the target to provide a high detection efficiency even in the presence of a distractor feature. This work therefore highlights the effectiveness of combining neuromorphic photonic hardware and software SNNs, for efficient learning and ultrafast operation, thanks to the use of spiking light signals, towards tackling complex AI and computer vision problems.

Research for the Business Role and Value of AI in the Medical Industry

This paper discusses the development of AI in healthcare, from the past to the present, and the future. We carefully analyzed the process of AI penetration into the medical field and also envisioned the future development of AI in the medical field. Because of AI's excellent computing power and analytical capabilities, it has a great impact on the medical industry. This paper describes the use and value of AI and sorts out the relationship between advantages and disadvantages from the aspects of image diagnosis, pharmacy, ChatGPT and human-computer interaction equipment. Based on recent research, the position of AI in the medical field is demonstrated from multiple perspectives, and their future market is speculated, for various reasons, some development cycles are relatively long, and many aspects need to be considered. Finally, this article presents some of the existing problems of AI in healthcare, from different aspects of patients, hospitals and doctors, and then gives different solutions.

AI Development in Germany After the Publication of German Nation AI Strategy, Problems and Solutions

Germany’s effort to develop AI and make progress in this field became coherent and systematic after the publication of the German National AI Strategy in 2018. Germany’s effort in AI development has four trends, namely the integration of AI in society and economy, building and maintaining the AI ecosystem, strengthening the country in terms of infrastructure and data, and finally research and development, which constitutes the essence and backbone of all the other trends. To drive its efforts forward in each trend, Germany should overcome the respective challenges. Furthermore, due to the novelty of the subject of AI and its fast-developing nature, AI has posed some challenges and issues in legal, political, and social fields that need to be addressed swiftly. Germans dominantly address these challenges either on national or European levels and mainly resort to research and academia, their mighty industrial sector, and legislation on federal and European levels to do so.

SOCIAL AND PHILOSOPHICAL FOUNDATIONS OF CURRENT PROBLEMS OF ARTIFICIAL INTELLIGENCE

The development of artificial intelligence technologies is an important condition for the formation of a digital society. However, according to a number of AI developers and researchers, this process can lead to catastrophic changes in the social, economic, and political spheres. The article discusses modern points of view on the risks of AI development: analyses the opinions of AI developers about the possible risks of technology development, describes the directions of philosophical research in AI, analyses forecasts of the social consequences of technology development. The scientific novelty lies in the integrated approach to considering the problem, based on the analysis of the social and ethical risks of AI. The article shows that at the moment the main means of preventing the risks of AI is to improve the system of its ethical and legal regulation. As a result, conclusions are drawn about the possible prospects for the development of the most complex problems of AI in the socio-philosophical context.

Active Predictive Coding: A Unifying Neural Model for Active Perception, Compositional Learning, and Hierarchical Planning.

There is growing interest in predictive coding as a model of how the brain learns through predictions and prediction errors. Predictive coding models have traditionally focused on sensory coding and perception. Here we introduce active predictive coding (APC) as a unifying model for perception, action, and cognition. The APC model addresses important open problems in cognitive science and AI, including (1) how we learn compositional representations (e.g., part-whole hierarchies for equivariant vision) and (2) how we solve large-scale planning problems, which are hard for traditional reinforcement learning, by composing complex state dynamics and abstract actions from simpler dynamics and primitive actions. By using hypernetworks, self-supervised learning, and reinforcement learning, APC learns hierarchical world models by combining task-invariant state transition networks and task-dependent policy networks at multiple abstraction levels. We illustrate the applicability of the APC model to active visual perception and hierarchical planning. Our results represent, to our knowledge, the first proof-of-concept demonstration of a unified approach to addressing the part-whole learning problem in vision, the nested reference frames learning problem in cognition, and the integrated state-action hierarchy learning problem in reinforcement learning.

Saturation-Based Boolean Conjunctive Query Answering and Rewriting for the Guarded Quantification Fragments

Query answering is an important problem in AI, database and knowledge representation. In this paper, we develop saturation-based Boolean conjunctive query answering and rewriting procedures for the guarded, the loosely guarded and the clique-guarded fragments. Our query answering procedure improves existing resolution-based decision procedures for the guarded and the loosely guarded fragments and this procedure solves Boolean conjunctive query answering problems for the guarded, the loosely guarded and the clique-guarded fragments. Based on this query answering procedure, we also introduce a novel saturation-based query rewriting procedure for these guarded fragments. Unlike mainstream query answering and rewriting methods, our procedures derive a compact and reusable saturation, namely a closure of formulas, to handle the challenge of querying for distributed datasets. This paper lays the theoretical foundations for the first automated deduction decision procedures for Boolean conjunctive query answering and the first saturation-based Boolean conjunctive query rewriting in the guarded, the loosely guarded and the clique-guarded fragments.

Exploring Applications and User Experience with Generative AI Tools: A Content Analysis of Reddit Posts on ChatGPT

ABSTRACTAs part of a larger project, this paper reports on preliminary findings of a study exploring use cases of ChatGPT and associated behaviors and experiences among users of an online forum. Posts on a ChatGPT‐related forum on Reddit (n = 452) were analyzed using qualitative content analysis. This paper reports on themes relevant to this study, including the types of tasks for which users used ChatGPT, user experiences, and perceived impacts of ChatGPT. ChatGPT was often used to facilitate various writing tasks (e.g., writing an essay), academic tasks (e.g., finding scientific references for a research paper), everyday tasks (e.g., creating a meal plan), and conversational purposes (e.g., having a simulated conversation about a past event). Users expressed positive (e.g., excited, amazed) and negative (e.g., fooled, concerned) feelings toward the technology. They raised various issues and problems with ChatGPT at the content (e.g., inaccuracy, incompletes) and system (e.g., unavailability, instability) levels. Users discussed the perceived impacts of ChatGPT on individuals (e.g., unemployment) and society (e.g., AI divide). Study findings can inform the design of policies and guidelines for mitigating AI problems and promoting the effective and ethical use of emerging AI technologies.