Tekst jest zapisem wywiadu, jaki z autorem przeprowadziła red. Katarzyna Kachel. Przedmiotem wywiadu były możliwości i ograniczenia sztucznej inteligencji. Omówiono różne zastosowania i sukcesy sztucznej inteligencji, ale podjęto także... more
Tekst jest zapisem wywiadu, jaki z autorem przeprowadziła red. Katarzyna Kachel. Przedmiotem wywiadu były możliwości i ograniczenia sztucznej inteligencji. Omówiono różne zastosowania i sukcesy sztucznej inteligencji, ale podjęto także bardzo rzeczową dyskusję na temat potencjalnych zagrożeń rozwoju sztucznej inteligencji.
Chương trình phát hiện khuôn mặt hoạt động như thế nào? (Sử dụng mạng thần kinh) Gần đây, tôi đã thử nghiệm với mô hình Mạng lưới kết nối phân tầng đa tác vụ (MTCNN) để nhận diện khuôn mặt. Mô hình này có ba mạng phức hợp (P-Net, R-Net và... more
Chương trình phát hiện khuôn mặt hoạt động như thế nào? (Sử dụng mạng thần kinh) Gần đây, tôi đã thử nghiệm với mô hình Mạng lưới kết nối phân tầng đa tác vụ (MTCNN) để nhận diện khuôn mặt. Mô hình này có ba mạng phức hợp (P-Net, R-Net và O-Net) và có thể hoạt động tốt hơn nhiều điểm chuẩn nhận diện khuôn mặt trong khi vẫn giữ được hiệu suất thời gian thực. Nhưng làm thế nào, chính xác, nó hoạt động như thế nào? Lưu ý: Nếu bạn muốn có một ví dụ cụ thể về cách xử lý mạng nơ-ron nhận diện khuôn mặt, tôi đã đính kèm các liên kết tải xuống của mô hình MTCNN bên dưới. Sau khi tải xuống, hãy mở ./mtcnn/mtcnn.py và cuộn đến chức năng detector_faces. Đây là hàm mà bạn sẽ gọi khi thực hiện mô hình này, vì vậy việc xem qua hàm này sẽ cho bạn biết cách chương trình tính toán và thu hẹp tọa độ của hộp giới hạn và các đặc điểm trên khuôn mặt. Tuy nhiên, tôi sẽ không giải mã từng dòng mã: nó không chỉ làm cho bài viết này dài một cách không cần thiết mà còn phản tác dụng đối với hầu hết người đọc vì rất nhiều tham số trong mã chỉ phù hợp với mô hình cụ thể này. Giai đoạn 1:
This monograph relates the concepts of computer intelligence and computer emotions to foreign language learning by humans. The discussion touches on two opposing theoretical approaches to Artificial Intelligence, one in favor of computer... more
This monograph relates the concepts of computer intelligence and computer emotions to foreign language learning by humans. The discussion touches on two opposing theoretical approaches to Artificial Intelligence, one in favor of computer intelligence, the other against it. Alan Turing (1950) offered up philosophical "proof" that computers can think. On the other hand, John Searle's (1980) Chinese Room analogy offered "proof" that computers cannot think. More recently, Umberto Eco (2003) discussed specific problems in translating from one language to another via a computer program. When translating back into the original language, the results are never the same as in the original. Eco observed that the task requires more than a simple knowledge of grammar. While most languages possess a larger number of meanings than words, machines are incapable of distinguishing among the multiple meanings of individual words.
This paper was presented at the International Conference of Education, Social Sciences, etc, San Francisco, 2016. It comprises an update of my previous paper, entitled HAL vs. Poole, 2001 -- Artificial Intelligence and Foreign Language Learning (2008).
Medical diagnosis involves a complex decision process that involves a lot of vagueness and uncertainty management, especially when the disease has multiple symptoms. A number of researchers have utilized the fuzzy-analytical hierarchy... more
Medical diagnosis involves a complex decision process that involves a lot of vagueness and uncertainty management, especially when the disease has multiple symptoms. A number of researchers have utilized the fuzzy-analytical hierarchy process (fuzzy-AHP) methodology in handling imprecise data in medical diagnosis and therapy. This is because fuzzy-AHP system is capable of accommodating inherent uncertainty and vagueness in multi-criteria decision making with hierarchical structuring. This study attempts to do a case comparison of the effectiveness of the fuzzy verses the AHP methodology in medical diagnosis in order to provide a framework for determining the appropriate backbone in a fuzzy-AHP hybrid system. The results of the study indicate a non-significant relative superiority of the fuzzy technology over the AHP technology.
This research investigated the use of artificial intelligence (AI) and machine learning in health tech startups to manage coronavirus. The study uses the case study approach to provide valuable insights showing that the health tech... more
This research investigated the use of artificial intelligence (AI) and machine learning in health tech startups to manage coronavirus. The study uses the case study approach to provide valuable insights showing that the health tech start-ups are helping the community to fight the pandemic. This study established that certain development strategies such as app-based solution for the healthcare information, the Unstructured Supplementary Service Data (USSD), and the use of geo-mobility intelligence in controlling the spread of COVID-19 mainly in Rural India have proved to be successful. This study also highlights some of the health tech start-ups' (Aiisma, Qure.ai, and TruFactor) initiatives that are helping the community and health industry. These start-ups mainly made their focus on the use of AI, and machine learning that have been chosen to help in the examination of the current needs, use, and roles of technology in controlling the COVID-19 pandemic.
Przedstawiony tekst jest zapisem wywiadu, jaki z autorem przeprowadził Andrzej Politowicz z Uniwersytetu w Zielonej Górze. Wywiad ten był przeprowadzony z okazji nadania autorowi tytułu doktora honoris causa Uniwersytetu Zielonogórskiego... more
Przedstawiony tekst jest zapisem wywiadu, jaki z autorem przeprowadził Andrzej Politowicz z Uniwersytetu w Zielonej Górze. Wywiad ten był przeprowadzony z okazji nadania autorowi tytułu doktora honoris causa Uniwersytetu Zielonogórskiego i był publikowany pierwotnie w Miesięczniku Społeczności Akademickiej tego Uniwersytetu. Wątek DHC jest jednak tylko wzmiankowany w tekście, natomiast zasadnicza treść dotyczy wybranych zagadnień sztucznej inteligencji. Ich omówienie stanowi główną treść artykułu, który został opublikowany także w Biuletynie Informacyjnym Pracowników AGH – i ta właśnie wersja jest tu eksponowana. Godny uwagi jest podtytuł tego tekstu: „Mózg nie męczy się od myślenia”
As data analysis tasks often have to face the analysis of huge and complex data sets there is a need for new algorithms that combine vector quantization and mapping methods to visualize the hidden data structure in a low-dimensional... more
As data analysis tasks often have to face the analysis of huge and complex data sets there is a need for new algorithms that combine vector quantization and mapping methods to visualize the hidden data structure in a low-dimensional vector space. In this paper a new class of algorithms is defined. Topology representing networks are applied to quantify and disclose the data structure and different nonlinear mapping algorithms for the low dimensional visualization are applied for the mapping of the quantized data. To evaluate the main properties of the resulted topology representing network based mapping methods a detailed analysis based on the wine benchmark example is given.
An emerging technique, inspired from the natural and social tendency of individuals to learn from each other referred to as Cohort Intelligence (CI) is presented. Learning here refers to a cohort candidate’s effort to self supervise its... more
An emerging technique, inspired from the natural and social tendency of individuals to learn from each other referred to as Cohort Intelligence (CI) is presented. Learning here refers to a cohort candidate’s effort to self supervise its own behavior and further adapt to the behavior of the other candidate which it intends to follow. This makes every candidate improve/evolve its behavior and eventually the entire cohort behavior. This ability of the approach is tested by solving an NP-hard combinatorial problem such as Knapsack Problem (KP). Several cases of the 0–1 KP are solved. The effect of various parameters on the solution quality has been discussed.The advantages and limitations of the CI methodology are also discussed.
Many sentient beings suffer serious harms due to a lack of moral consideration. Importantly, such harms could also occur to a potentially astronomical number of morally considerable future beings. This paper argues that, to prevent such... more
Many sentient beings suffer serious harms due to a lack of moral consideration. Importantly, such harms could also occur to a potentially astronomical number of morally considerable future beings. This paper argues that, to prevent such existential risks, we should prioritise the strategy of expanding humanity’s moral circle to include, ideally, all sentient beings. We present empirical evidence that, at micro- and macro-levels of society, increased concern for members of some outlying groups facilitates concern for others. We argue that the perspective of moral circle expansion can reveal and clarify important issues in futures studies, particularly regarding animal ethics and artificial intelligence. While the case for moral circle expansion does not hinge on specific moral criteria, we focus on sentience as the most recommendable policy when deciding, as we do, under moral uncertainty. We also address various nuances of adjusting the moral circle, such as the risk of over-expansion.
Industrial automation is not possible without artificial intelligence and machine learning. The goal of this study was to develop, verify, validate and train discriminant models for identification and classification of sorghum varieties... more
Industrial automation is not possible without artificial intelligence and machine learning. The goal of this study was to develop, verify, validate and train discriminant models for identification and classification of sorghum varieties using its optical properties. Three varieties of sorghum considered are NGB 01907 (Red Sorghum), NGB 01589 (White Sorghum) and NGB 01227 (Yellow Sorghum).Optical properties studied are colour (L a b), absorbance, reflectance and transmittance. Colour was measured using Chroma Meter, absorbance and transmittance were determined using spectrophotometer. Reflectance was calculated using Beer-Lambert equation. The experimental results shows that only about 15% of the light rays that reached the grains were absorbed, transmitted and directly reflected and almost 95% of the light rays shatteringly reflected. Two discriminant models were developed, the first one was found to be stronger than the second one. The colour properties (L a b) were found to have a greater impact on the models ability to discriminate than the other properties. Territorial map was developed that shows boundaries and regions which demarcate varieties or groups membership. The models predictive capacities were trained using the experimental data and the discriminant models verified and classified sorghum correctly for all varieties. The leave one cross validation method was used and the discriminant models also identify and classify 100% of all varieties group membership. The models ability was proven to predict both within and outside the experimental range of this study.
