Figure 1 - uploaded by Sebastian Schuchmann
Content may be subject to copyright.
Source publication
![Figure 2: The Mark 1 Perceptron [24]](profile/Sebastian-Schuchmann/publication/333039347/figure/fig2/AS:757796412858369@1557684152167/The-Mark-1-Perceptron-24_Q320.jpg)
![Figure 3: Data from 16,625 papers from arXiv. [41]](profile/Sebastian-Schuchmann/publication/333039347/figure/fig3/AS:757796412866560@1557684152214/Data-from-16-625-papers-from-arXiv-41_Q320.jpg)
![Figure 4: Gartner's Hype Cycle. [68]](profile/Sebastian-Schuchmann/publication/333039347/figure/fig4/AS:757796412854272@1557684152966/Gartners-Hype-Cycle-68_Q320.jpg)
This paper examines the probability of an approaching AI winter by comparing historical patterns to the modern era. First, the historical literature is analyzed and the central causes leading to an AI winter period are distilled and discussed:
- Expectations and promises compared to the actual results;
- Funding from governments and industries;
-...
Similar publications
Citations
... Despite early critiques of conversational AI and its societal implications, the timeline continues with further experiments building upon the technological foundation set by ELIZA, laying the groundwork for the chronological progression of conversational AI, such as PARRY in 1972, designed to simulate a patient with schizophrenia, thus reflecting the counterpart to Freud's divan (Adamopoulou & Moussiades, 2020). The chronology extends into periods of reduced research activity, often referred to as the AI Winter, which was marked by a notable decline in field in terms of funding and research (Schuchmann, 2019). Within this technological progression, key milestones took place at the end of the twentieth century. ...
The development and deployment of chatbot technology, while spanning decades and employing different techniques, require innovative frameworks to understand and interrogate their functionality and implications. A mere technocentric account of the evolution of chatbot technology does not fully illuminate how conversational systems are embedded in societal dynamics. This study presents a structured examination of chatbots across three societal dimensions, highlighting their roles as objects of scientific research, commercial instruments, and agents of intimate interaction. Through furnishing a dimensional framework for the evolution of conversational systems — from laboratories to marketplaces to private lives— this article contributes to the wider scholarly inquiry of chatbot technology and its impact in lived human experiences and dynamics.
... Over the past few decades we observed rises and falls of various trends and methods. There are periods of low interest in AI known as AI winter (Schuchmann, 2019). Starting around 2012 we could observe the beginning of a new method known as deep learning that quickly caught the attention of researchers and industry. ...
In the paper, we propose two models of Artificial Intelligence (AI) patents in European Union (EU) countries addressing spatial and temporal behaviour. In particular, the models can quantitatively describe the interaction between countries or explain the rapidly growing trends in AI patents. For spatial analysis Poisson regression is used to explain collaboration between a pair of countries measured by the number of common patents. Through Bayesian inference, we estimated the strengths of interactions between countries in the EU and the rest of the world. In particular, a significant lack of cooperation has been identified for some pairs of countries. Alternatively, an inhomogeneous Poisson process combined with the logistic curve growth accurately models the temporal behaviour by an accurate trend line. Bayesian analysis in the time domain revealed an upcoming slowdown in patenting intensity.
... Over the past few decades we observed raises and falls of various trends and methods. There are periods of low interests in AI known as AI winter Schuchmann (2019). Starting around 2012 we could observe the beginning of a new method known as deep learning that quickly caught the attention of the researchers and the industry. ...
In the paper, we propose two models of Artificial Intelligence (AI) patents in European Union (EU) countries addressing spatial and temporal behaviour. In particular, the models can quantitatively describe the interaction between countries or explain the rapidly growing trends in AI patents. For spatial analysis Poisson regression is used to explain collaboration between a pair of countries measured by the number of common patents. Through Bayesian inference, we estimated the strengths of interactions between countries in the EU and the rest of the world. In particular, a significant lack of cooperation has been identified for some pairs of countries. Alternatively, an inhomogeneous Poisson process combined with the logistic curve growth accurately models the temporal behaviour by an accurate trend line. Bayesian analysis in the time domain revealed an upcoming slowdown in patenting intensity.
... Expert systems are computer programs that aim to model human expertise to solve complex problems in one or more specific knowledge areas [54]. The field of AI experienced another major winter from 1987 to 1993 [55]. This was because expert systems computers were proven to be clumsy and slow as compared to desktop computers built by Apple and IBM [53]. ...
Solar energy powered systems are increasingly being implemented in different areas due to the advances in solar energy technologies. Some of the major areas for solar energy applications include solar water heating, solar electric power generation, and solar water pumping. Solar water pumping has become the most adopted solar energy technology in the last decade. It has been considered as an attractive way to provide water in remote areas. A major advantage of using solar water pumps is that they are naturally matched with solar irradiation since usually water demand is high in summer when solar irradiation has its maximum values. However, solar energy powered systems are weather dependent. In most cases, a solar energy source has to be combined with another energy source to form a hybrid system to overcome the demerits of using solar alone.
This thesis provides the detailed design, modelling and analysis of an Artificial Intelligence (AI) based solar/diesel hybrid water pumping system. This research aims to develop an optimization model that uses AI techniques to maximize the solar energy output and manage the energy flow within the solar/diesel hybrid water pumping. Thus, the proposed system is composed of solar photovoltaic modules, battery bank, Variable Speed Diesel Generator (VSDG), Adaptive Neuro-Fuzzy Inference System (ANFIS) based Maximum Power Point Tracking (MPPT) controllers and an Energy Management Controller (EMC). The EMC, which is based on Fuzzy Logic (FL), is responsible for managing the flow of energy throughout the hybrid system to ensure an undisturbed power supply to the water pump. The PV array, battery bank, VSDG are all sized to power a 5Hp DC water pump and the ANFIS based MPPT controllers are proposed for improving the efficiency of PV modules.
The modelling of the system components is performed in the MATLAB/Simulink environment. For evaluation of the proposed system, several case scenarios were considered and simulated in the MATLAB/Simulink environment. The simulation results revealed the effectiveness of the proposed ANFIS based MPPT controllers since the controllers were able to extract maximum available power from PV modules for both steady-state and varying weather conditions. The proposed EMC demonstrated the successful management and control of the energy flow within the hybrid system with less dependency on the VSDG. The EMC was also able to regulate the charging and discharging of the battery bank.
... NNs are one of the multiple AI techniques, and they have already been proposed for the physical layer. Coinciding with the AI spring in the 90s [17], NNs were proposed for channel estimation and equalization. Examples can be found in [18], [19] and [20]. ...
... Regardless of how the ML community decides to approach the architectural integration, it seems clear that, in the long term, AutonoML has the potential to revolutionise the way that ML is practiced, with significant implications for everyday decision-making across diverse facets of the human experience. That said, history has also witnessed at least two AI winters, primarily on the back of unmet expectations, and there is debate whether the limitations of deep learning are severe enough to herald another in the near future [308]. Thus, while crucial advances in HPO, meta-learning, NAS, adaptation, and so on, are arguably academically impressive, the evolution of the field is not served by overhype. ...
Over the last decade, the long-running endeavour to automate high-level processes in machine learning (ML) has risen to mainstream prominence, stimulated by advances in optimisation techniques and their impact on selecting ML models/algorithms. Central to this drive is the appeal of engineering a computational system that both discovers and deploys high-performance solutions to arbitrary ML problems with minimal human interaction. Beyond this, an even loftier goal is the pursuit of autonomy, which describes the capability of the system to independently adjust an ML solution over a lifetime of changing contexts. However, these ambitions are unlikely to be achieved in a robust manner without the broader synthesis of various mechanisms and theoretical frameworks, which, at the present time, remain scattered across numerous research threads. Accordingly, this review seeks to motivate a more expansive perspective on what constitutes an automated/autonomous ML system, alongside consideration of how best to consolidate those elements. In doing so, we survey developments in the following research areas: hyperparameter optimisation, multi-component models, neural architecture search, automated feature engineering, meta-learning, multi-level ensembling, dynamic adaptation, multi-objective evaluation, resource constraints, flexible user involvement, and the principles of generalisation. We also develop a conceptual framework throughout the review, augmented by each topic, to illustrate one possible way of fusing high-level mechanisms into an autonomous ML system. Ultimately, we conclude that the notion of architectural integration deserves more discussion, without which the field of automated ML risks stifling both its technical advantages and general uptake.
Apart from being a buzzword, AI is a rapidly developing, complex, interdisciplinary research area. Seen as a technology capable of learning from experience and acting autonomously without human intervention, AI can be one of the, if not the most disruptive and transformative piece of technology developed by humanity. To date, AI applications can be found in a vast spectrum of available technologies, starting from consumer appliances and up to fully autonomous vehicles. The impact of AI can be felt in various areas, such as financial markets, state administration, healthcare, transportation, physical and digital infrastructure. The global economy at large is benefiting vastly from the advancements in AI research and development. According to some estimates, AI could potentially create 3.5 trillion US Dollars (USD) to 5.8 trillion USD in annual value in the global economy. Moreover, considering significant investments made by leading tech corporations worldwide to the research and development of AI, companies’ total AI absorption level is estimated to reach about 50 per cent by the year 2030.
