Human Collective Intelligence Research Articles

Evaluating multimodal AI in medical diagnostics

This study evaluates multimodal AI models’ accuracy and responsiveness in answering NEJM Image Challenge questions, juxtaposed with human collective intelligence, underscoring AI’s potential and current limitations in clinical diagnostics. Anthropic’s Claude 3 family demonstrated the highest accuracy among the evaluated AI models, surpassing the average human accuracy, while collective human decision-making outperformed all AI models. GPT-4 Vision Preview exhibited selectivity, responding more to easier questions with smaller images and longer questions.

Augmenting fake content detection in online platforms: A domain adaptive transfer learning via adversarial training approach

Online platforms are experimenting with interventions such as content screening to moderate the effects of fake, biased, and incensing content. Yet, online platforms face an operational challenge in implementing machine learning algorithms for managing online content due to the labeling problem, where labeled data used for model training are limited and costly to obtain. To address this issue, we propose a domain adaptive transfer learning via adversarial training approach to augment fake content detection with collective human intelligence. We first start with a source domain dataset containing deceptive and trustworthy general news constructed from a large collection of labeled news sources based on human judgments and opinions. We then extract discriminating linguistic features commonly found in source domain news using advanced deep learning models. We transfer these features associated with the source domain to augment fake content detection in three target domains: political news, financial news, and online reviews. We show that domain invariant linguistic features learned from a source domain with abundant labeled examples can effectively improve fake content detection in a target domain with very few or highly unbalanced labeled data. We further show that these linguistic features offer the most value when the level of transferability between source and target domains is relatively high. Our study sheds light on the platform operation in managing online content and resources when applying machine learning for fake content detection. We also outline a modular architecture that can be adopted in developing content screening tools in a wide spectrum of fields.

Human Collective Intelligence Inspired Multi-View Representation Learning - Enabling View Communication by Simulating Human Communication Mechanism.

In real-world applications, we often encounter multi-view learning tasks where we need to learn from multiple sources of data or use multiple sources of data to make decisions. Multi-view representation learning, which can learn a unified representation from multiple data sources, is a key pre-task of multi-view learning and plays a significant role in real-world applications. Accordingly, how to improve the performance of multi-view representation learning is an important issue. In this work, inspired by human collective intelligence shown in group decision making, we introduce the concept of view communication into multi-view representation learning. Furthermore, by simulating human communication mechanism, we propose a novel multi-view representation learning approach that can fulfill multi-round view communication. Thus, each view of our approach can exploit the complementary information from other views to help with modeling its own representation, and mutual help between views is achieved. Extensive experiment results on six datasets from three significant fields indicate that our approach substantially improves the average classification accuracy by 4.536% in medicine and bioinformatics fields as well as 4.115% in machine learning field.

Toward a Superintelligent Action Recommender for Network Operation Centers Using Reinforcement Learning

Today’s Network Operation Centres (NOC) consist of teams of network professionals responsible for monitoring and taking actions for their network’s health. Most of these NOC actions are relatively complex and executed manually; only the simplest tasks can be automated with rules-based software. But today’s networks are getting larger and more complex. Therefore, deciding what action to take in the face of non-trivial problems has essentially become an art that depends on collective human intelligence of NOC technicians, specialized support teams organized by technology domains, and vendors’ technical support. This model is getting increasingly expensive and inefficient, and the automation of all or at least some NOC tasks is now considered a desirable step towards autonomous and self-healing networks. In this article, we investigate whether an autonomous NOC can achieve superintelligence; i.e., it can recommend or take actions that lead to better results than those achieved by the collective human intelligence, in this case rules designed by human experts. Our investigation is inspired by the superintelligence achieved in computer games recently. Specifically, we build an Action Recommendation Engine based on Reinforcement Learning, a type of Artificial Intelligence method, and we train it with expert rules and lets it explore actions by itself. We then show that it can learn new and more efficient strategies that outperform expert rules designed by humans. This can be used in face of network problems to either quickly recommend actions to NOC technicians or autonomously take actions for fast recovery.

Society 5.0: Internet as if People Mattered

While the primary focus of 5G has been on industry verticals, future 6G mobile networks are anticipated to become more human-centered. Emerging cyber-physical-social systems (CPSSs) aim at functionally integrating human beings into today's cyber-physical systems at the social, cognitive, and physical levels. CPSSs are instrumental in realizing the human-centered Society 5.0 vision. Society 5.0 envisions human beings increasingly interacting with social robots and embodied artificial intelligence in their daily lives. In this article, we build on our recent work on robonomics in the 6G era. Robonomics is an emerging field that investigates social human-robot interaction and its sociotechnical impact as well as blockchain technologies and cryptocurrencies, not only coins but - more interestingly - also tokens. Specifically, we study the tokenization process of creating tokenized digital twins of assets and access rights in the physical and digital world, paying close attention to its central role in ushering in the future Web3 and its underlying token economy, the successor of today's information and platform economies. After introducing our CPSS-based bottom-up multilayer token engineering framework for Society 5.0, we experimentally demonstrate how the collective human intelligence of a blockchain-enabled decentralized autonomous organization can be enhanced via purpose-driven tokens.

The network science of collective intelligence.

In the last few years, breakthroughs in computational and experimental techniques have produced several key discoveries in the science of networks and human collective intelligence. This review presents the latest scientific findings from two key fields of research: collective problem-solving and the wisdom of the crowd. I demonstrate the core theoretical tensions separating these research traditions and show how recent findings offer a new synthesis for understanding how network dynamics alter collective intelligence, both positively and negatively. I conclude by highlighting current theoretical problems at the forefront of research on networked collective intelligence, as well as vital public policy challenges that require new research efforts.

Formal verification of persistence and liveness in the trust-based blockchain crowdsourcing consensus protocol

Crowdsourcing is a potential computing paradigm that exploits collective human intelligence to solve complex tasks, but it suffers from various safety and security problems. Blockchain has emerged as a promising technology to address most of the security issues, however, it is challenging to find an appropriate and trusted blockchain-based consensus protocol for crowdsourcing services. This work proposes a novel Trust-based Blockchain Crowdsourcing consensus protocol that selects a leader and validators based on various trust factors. The proposed protocol addresses a major issue of ensuring correctness associated with the safety and security-critical systems which has a vital importance because failure of such systems may lead to adverse consequences. Mainly it is focused on persistence and liveness properties preventing invalid block insertion and consensus delay attacks. Model checking technique is utilized because of its effectiveness and automatic nature to perform formal verification. The proposed protocol is specified using Communicating Sequential Programs, and the persistence and liveness properties are specified through Linear Temporal Logic. The model verification is performed by giving the formal model and the properties as input to the Process Analysis Toolkit which checks for the satisfaction or violation of the properties.

The origins of human cumulative culture: from the foraging niche to collective intelligence.

Various studies have investigated cognitive mechanisms underlying culture in humans and other great apes. However, the adaptive reasons for the evolution of uniquely sophisticated cumulative culture in our species remain unclear. We propose that the cultural capabilities of humans are the evolutionary result of a stepwise transition from the ape-like lifestyle of earlier hominins to the foraging niche still observed in extant hunter–gatherers. Recent ethnographic, archaeological and genetic studies have provided compelling evidence that the components of the foraging niche (social egalitarianism, sexual and social division of labour, extensive co-residence and cooperation with unrelated individuals, multilocality, fluid sociality and high between-camp mobility) engendered a unique multilevel social structure where the cognitive mechanisms underlying cultural evolution (high-fidelity transmission, innovation, teaching, recombination, ratcheting) evolved as adaptations. Therefore, multilevel sociality underlies a ‘social ratchet’ or irreversible task specialization splitting the burden of cultural knowledge across individuals, which may explain why human collective intelligence is uniquely able to produce sophisticated cumulative culture. The foraging niche perspective may explain why a complex gene-culture dual inheritance system evolved uniquely in humans and interprets the cultural, morphological and genetic origins of Homo sapiens as a process of recombination of innovations appearing in differentiated but interconnected populations.This article is part of a discussion meeting issue ‘The emergence of collective knowledge and cumulative culture in animals, humans and machines’.

Individual and collective human intelligence in drug design: evaluating the search strategy

In recent years, individual and collective human intelligence, defined as the knowledge, skills, reasoning and intuition of individuals and groups, have been used in combination with computer algorithms to solve complex scientific problems. Such approach was successfully used in different research fields such as: structural biology, comparative genomics, macromolecular crystallography and RNA design. Herein we describe an attempt to use a similar approach in small-molecule drug discovery, specifically to drive search strategies of de novo drug design. This is assessed with a case study that consists of a series of public experiments in which participants had to explore the huge chemical space in silico to find predefined compounds by designing molecules and analyzing the score associate with them. Such a process may be seen as an instantaneous surrogate of the classical design-make-test cycles carried out by medicinal chemists during the drug discovery hit to lead phase but not hindered by long synthesis and testing times. We present first findings on (1) assessing human intelligence in chemical space exploration, (2) comparing individual and collective human intelligence performance in this task and (3) contrasting some human and artificial intelligence achievements in de novo drug design.

Collective human intelligence outperforms artificial intelligence in a skin lesion classification task.

Convolutional neural networks (CNN) enable accurate diagnosis of medical images and perform on or above the level of individual physicians. Recently, collective human intelligence (CoHI) was shown to exceed the diagnostic accuracy of individuals. Thus, diagnostic performance of CoHI (120 dermatologists) versus individual dermatologists versus two state-of-the-art CNN was investigated. Cross-sectional reader study with presentation of 30 clinical cases to 120 dermatologists. Six diagnoses were offered and votes collected via remote voting devices (quizzbox®, Quizzbox Solutions GmbH, Stuttgart, Germany). Dermatoscopic images were classified by a binary and multiclass CNN (FotoFinder Systems GmbH, Bad Birnbach, Germany). Three sets of diagnostic classifications were scored against ground truth: (1) CoHI, (2) individual dermatologists, and (3) CNN. CoHI attained a significantly higher accuracy [95% confidence interval] (80.0% [62.7%-90.5%]) than individual dermatologists (75.7% [73.8%-77.5%]) and CNN (70.0% [52.1%-83.3%]; all P<0.001) in binary classifications. Moreover, CoHI achieved a higher sensitivity (82.4% [59.0%-93.8%]) and specificity (76.9% [49.7%-91.8%]) than individual dermatologists (sensitivity 77.8% [75.3%-80.2%], specificity 73.0% [70.6%-75.4%]) and CNN (sensitivity 70.6% [46.9%-86.7%], specificity 69.2% [42.4%-87.3%]). The diagnostic accuracy of CoHI was superior to that of individual dermatologists (P < 0.001) in multiclass evaluation, with the accuracy of the latter comparable to multiclass CNN. Our analysis revealed that the majority vote of an interconnected group of dermatologists (CoHI) outperformed individuals and CNN in a demanding skin lesion classification task.

Solving pictorial jigsaw puzzles via Internet-based collective intelligence

Pictorial jigsaw (PJ) puzzles are a well-known leisure game for humans. Usually, a PJ puzzle game is played by one or several players face-to-face in physical space. In this paper, we focus on how to solve PJ puzzles in cyberspace by a group of physically-distributed players. We propose an approach to solve PJ puzzles using stigmergy-inspired Internet-based human collective intelligence. The core of the approach is a continuously-executing loop, namely, the EIFloop, which consists of three activities: exploration, integration, and feedback. During the explorationactivity, each player tries to solve the PJ puzzle alone, without direct interaction with the other players. At any time, the result of a players exploration represents a partial solution to the PJ puzzle and a set of rejected neighboring relationships between the puzzle pieces. The results of all players explorations are integrated in real time through the integrationactivity, forming a continuously-updated collective opinion graphCOG). Through the feedbackactivity, each player is provided with personalized feedback information based on the current COG and the players current exploration results, in order to accelerate his/her puzzle-solving process. Exploratory experiments show that:(1) Supported by this approach, the time to solve PJ puzzles is nearly linear with respect to the reciprocal of the number of players. Furthermore, compared with the best single players in the experiments, the puzzle-solving time decreases by 31.36%–64.57% on average for groups composed of 2 to 10 players.linebreak(2) Supported by this approach, the feedback information received by the best player in a group has an average precision of 86.34%, and, as the group size increases, the feedback information in the puzzle-solving result of the best player increases gradually from 20% to 45%.(3) This approach exhibits a better scalability with puzzle size than that of a face-to-face collaboration among ten players. Additionally, it always leads to 100%-accurate puzzle-solving results, whereas the results of the automated PJ-puzzle solver have an average accuracy of only 52%.It is envisaged that these results will provide useful information or opinions to facilitate a broader application of human collective intelligence in the Internet environment.

A primer of knowledge management for smart city governance

The concepts of smart city cannot be understood without examining their links with the knowledge society. In this kind of society, knowledge must be considered as a capital shared not only with experts, but also with citizens, within the target of sustainable development. After studying those concepts, the scope is to examine how human knowledge and artificial intelligence can be combined in geographic knowledge systems, essentially based on machine-processable knowledge and the concept of rules. Several geospatial rules are detailed in order to distinguish several categories implying locational issues. Then, a general structure for geographic knowledge base systems is given together with some fundamental elements for an infrastructure in which human collective intelligence is a key-element. To conclude this paper, a research agenda is given to integrate urban and regional management into the knowledge society.

About Territorial Intelligence and Geographic Knowledge Bases

As business intelligence is for companies, territorial intelligence is for local authorities. However, it is not the simple exportation of business intelligence concepts and tools, but overall a new discipline characterized by the amalgamation of collective human intelligence and computer intelligence. As a result, geographic knowledge bases, rules and inference engines are not only based on logic but also on topology, geometry and linguistics, and must take multiple actors’ points of view into account. The scope of this paper will be to present the concepts necessary to construct such systems, i.e. to pass from intelligence to geographic knowledge engineering.

HUMAN INTELLIGENCE OPERATED BY INNATE SIGN REFLEX LOGIC OF DICHOTOMY AND DUALISM AND POST-NATALLY ACQUIRED LOGIC OF “GROUP” AND “FORWARD ERROR CORRECTION (FEC)” - DIGITAL LINGUISTICS

It is plausible that language is processed by the innate spinal sign reflex mechanism. Human speech sound contains logical properties of phoneme and mora, which interact with in-brain antigen/antibody molecular structures and networking memories in the postnatally developed logical circuits optimized for signals of syllables, a character set (long lasting syllables) and electronic form (interactive and ubiquitous syllables). For grammatical decoding, sound localization function at the brainstem auditory nuclei is converted to process sound vectors of grammatical syllables. For axiomatic scientific concepts, sign reflex logic of “If A then B” must be upgraded by mathematical group logic to be acquired by learning, so that the results of interdisciplinary conceptual operations should become meaningful and productive. As electronic data is interactive, we can discover as many relevant linguistic information as possible with keyword search on the internet within a few seconds. As a result, we obtain multi-disciplinary and voluminous linguistic information but with uncertain reliability. To inherit collective human intelligence and correct any errors in linguistic information of the past, regenerative reading might be appropriate for individual ontogenesis to repeat intellectual and conceptual phylogeny. Errors cannot be corrected by the original authors posthumously but only by those who are alive now and equipped with error correction techniques for linguistic information.

Analysis of Collective Human Intelligence for Diagnosis of Pigmented Skin Lesions Harnessed by Gamification Via a Web-Based Training Platform: Simulation Reader Study.

BackgroundThe diagnosis of pigmented skin lesion is error prone and requires domain-specific expertise, which is not readily available in many parts of the world. Collective intelligence could potentially decrease the error rates of nonexperts.ObjectiveThe aim of this study was to evaluate the feasibility and impact of collective intelligence for the detection of skin cancer.MethodsWe created a gamified study platform on a stack of established Web technologies and presented 4216 dermatoscopic images of the most common benign and malignant pigmented skin lesions to 1245 human raters with different levels of experience. Raters were recruited via scientific meetings, mailing lists, and social media posts. Education was self-declared, and domain-specific experience was tested by screening tests. In the target test, the readers had to assign 30 dermatoscopic images to 1 of the 7 disease categories. The readers could repeat the test with different lesions at their own discretion. Collective human intelligence was achieved by sampling answers from multiple readers. The disease category with most votes was regarded as the collective vote per image.ResultsWe collected 111,019 single ratings, with a mean of 25.2 (SD 18.5) ratings per image. As single raters, nonexperts achieved a lower mean accuracy (58.6%) than experts (68.4%; mean difference=−9.4%; 95% CI −10.74% to −8.1%; P<.001). Collectives of nonexperts achieved higher accuracies than single raters, and the improvement increased with the size of the collective. A collective of 4 nonexperts surpassed single nonexperts in accuracy by 6.3% (95% CI 6.1% to 6.6%; P<.001). The accuracy of a collective of 8 nonexperts was 9.7% higher (95% CI 9.5% to 10.29%; P<.001) than that of single nonexperts, an improvement similar to single experts (P=.73). The sensitivity for malignant images increased for nonexperts (66.3% to 77.6%) and experts (64.6% to 79.4%) for answers given faster than the intrarater mean.ConclusionsA high number of raters can be attracted by elements of gamification and Web-based marketing via mailing lists and social media. Nonexperts increase their accuracy to expert level when acting as a collective, and faster answers correspond to higher accuracy. This information could be useful in a teledermatology setting.

An Information Theory of Language (Digital Linguistics)

Digital Linguistics (DL) is an interdisciplinary study that identifies human language as a digital evolution of mammal analog vocal sign communications, founded on the vertebrate spinal sign reflex mechanism. The author identifies the birth of linguistic humans at the time of the laryngeal descent, which provided vowel accented syllables containing logical properties of phonemes and morae. A character set and literacy is the second evolution. Written text is long lasting syllables in the brain of literate persons, and a civilization started as a linguistic phenomenon. At the end of their biological life, they can write their accumulated knowledge with a character set, so that subsequent generations can share their thoughts and experiences to develop further. In the 21st century, text is available in electronic form, which is interactive against keyword searches. It is time for linguistic humans to integrate disciplinary sciences and correct any errors to establish collective human intelligence. Piaget indicated that concepts can be manipulated correctly based on a group theory, which seems to be a base logic for intellectual evolution of linguistic humans as well as a tool for respectful and productive interdisciplinary discussions. [1][2]

Crowdsourcing Model for Multilingual Corpus and Knowledge Construction: The Case of Transnational Mark Twain

PURPOSE/THESIS: We describe a new approach that addresses key challenges to multilingual corpus by merging collective human intelligence (crowdsourcing) and automated knowledge construction and extraction methods in a symbiotic fashion. APPROACH/METHODS: We use a crowdsourcing model to collect and annotate translations of the same literary text. RESULTS AND CONCLUSIONS: The model promotes a dynamic approach to archives that increases the impact of traditional research by presenting the text from a new angle, accessible to a global public.PRACTICAL IMPLICATIONS: The Global Huck project proposes a new paradigm to assess the contribution of crowdsourcing-based models for collection and annotation purposes. ORIGINALITY/VALUE: Choosing the translations of a novel as a field of study is a truly transnational and multilingual collaborative endeavor allowing us to increase our capacity to collect and organize data on a broad, transnational and multilingual scale.

Online Academic Networks as Knowledge Brokers: The Mediating Role of Organizational Support

Placing online academic networks in the framework of social, cultural and institutional “deterritorialization,” the current paper aims at investigating the functionality of these new forms of transnational and trans-organizational aggregations as knowledge brokers. The emphasis is laid on the influence of human collective intelligence and consistent knowledge flows on research innovation, considering the role of organizational support within higher education systems. In this respect, the research relied on a questionnaire-based survey with 140 academics from European emerging countries, the data collected being processed via a partial least squares structural equation modelling technique. Evidence was brought that, as knowledge brokers, online academic networks are systems aimed to support the access to human collective intelligence and consistent knowledge flows which exert a positive influence on research innovation, both directly and indirectly, by means of formal and informal organizational support. As facilitators of collaborative environments for individuals with specialized knowledge, competence, expertise and experience, online academic networks have set themselves up as an agora for academics worldwide and as an outlet for their acumen and literacy.

Big Data Analytics on the Characteristic Equilibrium of Collective Opinions in Social Networks

Big data are products of human collective intelligence that are exponentially increasing in all facets of quantity, complexity, semantics, distribution, and processing costs in computer science, cognitive informatics, web-based computing, cloud computing, and computational intelligence. This paper presents fundamental big data analysis and mining technologies in the domain of social networks as a typical paradigm of big data engineering. A key principle of computational sociology known as the characteristic opinion equilibrium is revealed in social networks and electoral systems. A set of numerical and fuzzy models for collective opinion analyses is formally presented. Fuzzy data mining methodologies are rigorously described for collective opinion elicitation and benchmarking in order to enhance the conventional counting and statistical methodologies for big data analytics.

Wisdom of Artificial Crowds—A Metaheuristic Algorithm for Optimization

Finding optimal solutions to NP-Hard problems requires exponential time with respect to the size of the problem. Consequently, heuristic methods are usually utilized to obtain approximate solutions to problems of such difficulty. In this paper, a novel swarm-based nature-inspired metaheuristic algorithm for optimization is proposed. Inspired by human collective intelligence, Wisdom of Artificial Crowds (WoAC) algorithm relies on a group of simulated intelligent agents to arrive at independent solutions aggregated to produce a solution which in many cases is superior to individual solutions of all participating agents. We illustrate superior performance of WoAC by comparing it against another bio-inspired approach, the Genetic Algorithm, on one of the classical NP-Hard problems, the Travelling Salesperson Problem. On average a 3% - 10% improvement in quality of solutions is observed with little computational overhead.