ºÝºÝߣshows by User: canagnos / http://www.slideshare.net/images/logo.gif ºÝºÝߣshows by User: canagnos / Fri, 10 Aug 2018 07:13:50 GMT ºÝºÝߣShare feed for ºÝºÝߣshows by User: canagnos Weakly supervised learning /slideshow/weakly-supervised-learning/109326652 statisticalcyberisbis-180810071350
In many domains, the promise of disruptive innovation via the use of machine learning has not materialised due to the lack of the one resource that most machine learning models rely upon: labelled data. This is particularly true in domains where ground truth does not avail itself readily by way of a natural process or a business operation, but is rather the result of tedious manual labelling performed by human experts sifting through the data. One example is cybersecurity, where successful attacks are rare, and can go undetected for long periods of time, so that high-quality labels require a significant time investment from highly paid and exceptionally busy cyber analysts. Another example is natural language processing in niche areas that are awash with domain-specific jargon, such as police investigations. Such domains struggle to take advantage of the huge methodological and technological advances in supervised machine learning technology and are poorly served by most popular machine learning software packages. In this presentation, we take a step back to challenge the standard interface of “learning by example", and offer an alternative, more scalable way of incorporating expert opinion into a machine learning pipeline, known as weakly supervised learning. We discuss how this framework can be related to previous work, the new questions it poses, and the challenges and opportunities it presents us with from a technological perspective.]]>
In many domains, the promise of disruptive innovation via the use of machine learning has not materialised due to the lack of the one resource that most machine learning models rely upon: labelled data. This is particularly true in domains where ground truth does not avail itself readily by way of a natural process or a business operation, but is rather the result of tedious manual labelling performed by human experts sifting through the data. One example is cybersecurity, where successful attacks are rare, and can go undetected for long periods of time, so that high-quality labels require a significant time investment from highly paid and exceptionally busy cyber analysts. Another example is natural language processing in niche areas that are awash with domain-specific jargon, such as police investigations. Such domains struggle to take advantage of the huge methodological and technological advances in supervised machine learning technology and are poorly served by most popular machine learning software packages. In this presentation, we take a step back to challenge the standard interface of “learning by example", and offer an alternative, more scalable way of incorporating expert opinion into a machine learning pipeline, known as weakly supervised learning. We discuss how this framework can be related to previous work, the new questions it poses, and the challenges and opportunities it presents us with from a technological perspective.]]> Fri, 10 Aug 2018 07:13:50 GMT /slideshow/weakly-supervised-learning/109326652 canagnos@slideshare.net(canagnos) Weakly supervised learning canagnos In many domains, the promise of disruptive innovation via the use of machine learning has not materialised due to the lack of the one resource that most machine learning models rely upon: labelled data. This is particularly true in domains where ground truth does not avail itself readily by way of a natural process or a business operation, but is rather the result of tedious manual labelling performed by human experts sifting through the data. One example is cybersecurity, where successful attacks are rare, and can go undetected for long periods of time, so that high-quality labels require a significant time investment from highly paid and exceptionally busy cyber analysts. Another example is natural language processing in niche areas that are awash with domain-specific jargon, such as police investigations. Such domains struggle to take advantage of the huge methodological and technological advances in supervised machine learning technology and are poorly served by most popular machine learning software packages. In this presentation, we take a step back to challenge the standard interface of “learning by example", and offer an alternative, more scalable way of incorporating expert opinion into a machine learning pipeline, known as weakly supervised learning. We discuss how this framework can be related to previous work, the new questions it poses, and the challenges and opportunities it presents us with from a technological perspective. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/statisticalcyberisbis-180810071350-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> In many domains, the promise of disruptive innovation via the use of machine learning has not materialised due to the lack of the one resource that most machine learning models rely upon: labelled data. This is particularly true in domains where ground truth does not avail itself readily by way of a natural process or a business operation, but is rather the result of tedious manual labelling performed by human experts sifting through the data. One example is cybersecurity, where successful attacks are rare, and can go undetected for long periods of time, so that high-quality labels require a significant time investment from highly paid and exceptionally busy cyber analysts. Another example is natural language processing in niche areas that are awash with domain-specific jargon, such as police investigations. Such domains struggle to take advantage of the huge methodological and technological advances in supervised machine learning technology and are poorly served by most popular machine learning software packages. In this presentation, we take a step back to challenge the standard interface of “learning by example&quot;, and offer an alternative, more scalable way of incorporating expert opinion into a machine learning pipeline, known as weakly supervised learning. We discuss how this framework can be related to previous work, the new questions it poses, and the challenges and opportunities it presents us with from a technological perspective.
Weakly supervised learning from Christoforos Anagnostopoulos
]]> 305 2 https://cdn.slidesharecdn.com/ss_thumbnails/statisticalcyberisbis-180810071350-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Why Data Science is a Science /slideshow/why-data-science-is-a-science/109326276 athensdatasciencemeetup-180810071012
The term 'Data Scientist' arose fairly recently to express the specialised recruitment needs of certain well-known data-driven Silicon Valley firms. It signifies a mix of diverse and rare talents, mostly drawing from Computer Science (with emphasis on Big Data), Statistics and Machine Learning. In this talk, we will attempt to briefly survey the state-of-the-art both in terms of problems and solutions at the vanguard of Data Science. We will cover both novel developments, as well as centuries-old best practices, in an attempt to demonstrate that Data Science is indeed a Science, in the full sense of the word. This talk represents part of a seminar series that the speaker has given across the world, including Google (Mountainview), Cisco (San Jose) and Aviva Headquarters (London), and represents joint work with Professor David Hand (OBE). ]]>
The term 'Data Scientist' arose fairly recently to express the specialised recruitment needs of certain well-known data-driven Silicon Valley firms. It signifies a mix of diverse and rare talents, mostly drawing from Computer Science (with emphasis on Big Data), Statistics and Machine Learning. In this talk, we will attempt to briefly survey the state-of-the-art both in terms of problems and solutions at the vanguard of Data Science. We will cover both novel developments, as well as centuries-old best practices, in an attempt to demonstrate that Data Science is indeed a Science, in the full sense of the word. This talk represents part of a seminar series that the speaker has given across the world, including Google (Mountainview), Cisco (San Jose) and Aviva Headquarters (London), and represents joint work with Professor David Hand (OBE). ]]> Fri, 10 Aug 2018 07:10:12 GMT /slideshow/why-data-science-is-a-science/109326276 canagnos@slideshare.net(canagnos) Why Data Science is a Science canagnos The term 'Data Scientist' arose fairly recently to express the specialised recruitment needs of certain well-known data-driven Silicon Valley firms. It signifies a mix of diverse and rare talents, mostly drawing from Computer Science (with emphasis on Big Data), Statistics and Machine Learning. In this talk, we will attempt to briefly survey the state-of-the-art both in terms of problems and solutions at the vanguard of Data Science. We will cover both novel developments, as well as centuries-old best practices, in an attempt to demonstrate that Data Science is indeed a Science, in the full sense of the word. This talk represents part of a seminar series that the speaker has given across the world, including Google (Mountainview), Cisco (San Jose) and Aviva Headquarters (London), and represents joint work with Professor David Hand (OBE). <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/athensdatasciencemeetup-180810071012-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> The term &#39;Data Scientist&#39; arose fairly recently to express the specialised recruitment needs of certain well-known data-driven Silicon Valley firms. It signifies a mix of diverse and rare talents, mostly drawing from Computer Science (with emphasis on Big Data), Statistics and Machine Learning. In this talk, we will attempt to briefly survey the state-of-the-art both in terms of problems and solutions at the vanguard of Data Science. We will cover both novel developments, as well as centuries-old best practices, in an attempt to demonstrate that Data Science is indeed a Science, in the full sense of the word. This talk represents part of a seminar series that the speaker has given across the world, including Google (Mountainview), Cisco (San Jose) and Aviva Headquarters (London), and represents joint work with Professor David Hand (OBE).
Why Data Science is a Science from Christoforos Anagnostopoulos
]]> 93 3 https://cdn.slidesharecdn.com/ss_thumbnails/athensdatasciencemeetup-180810071012-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Data Science versus Artificial Intelligence: a useful distinction /slideshow/data-science-versus-artificial-intelligence-a-useful-distinction/78736791 mentatcanagnos-170810151953
A number of recent milestones in AI have rekindled the faith that human-grade computer intelligence can fuel the next technological revolution. In parallel and almost independently, the job role of Data Scientist rose to one of the hottest tickets in the technology sector. Despite the obvious overlap in the domains of Data Science and Artificial Intelligence, the two approaches are sufficiently distinct that choosing the wrong one might trigger a product to fail or a hiring process to go wrong. This presentation will offer some clarity and best practices with regards to understanding what data analysis requirements you really have, as what opposed to what you think you have.]]>
A number of recent milestones in AI have rekindled the faith that human-grade computer intelligence can fuel the next technological revolution. In parallel and almost independently, the job role of Data Scientist rose to one of the hottest tickets in the technology sector. Despite the obvious overlap in the domains of Data Science and Artificial Intelligence, the two approaches are sufficiently distinct that choosing the wrong one might trigger a product to fail or a hiring process to go wrong. This presentation will offer some clarity and best practices with regards to understanding what data analysis requirements you really have, as what opposed to what you think you have.]]> Thu, 10 Aug 2017 15:19:53 GMT /slideshow/data-science-versus-artificial-intelligence-a-useful-distinction/78736791 canagnos@slideshare.net(canagnos) Data Science versus Artificial Intelligence: a useful distinction canagnos A number of recent milestones in AI have rekindled the faith that human-grade computer intelligence can fuel the next technological revolution. In parallel and almost independently, the job role of Data Scientist rose to one of the hottest tickets in the technology sector. Despite the obvious overlap in the domains of Data Science and Artificial Intelligence, the two approaches are sufficiently distinct that choosing the wrong one might trigger a product to fail or a hiring process to go wrong. This presentation will offer some clarity and best practices with regards to understanding what data analysis requirements you really have, as what opposed to what you think you have. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/mentatcanagnos-170810151953-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> A number of recent milestones in AI have rekindled the faith that human-grade computer intelligence can fuel the next technological revolution. In parallel and almost independently, the job role of Data Scientist rose to one of the hottest tickets in the technology sector. Despite the obvious overlap in the domains of Data Science and Artificial Intelligence, the two approaches are sufficiently distinct that choosing the wrong one might trigger a product to fail or a hiring process to go wrong. This presentation will offer some clarity and best practices with regards to understanding what data analysis requirements you really have, as what opposed to what you think you have.
Data Science versus Artificial Intelligence: a useful distinction from Christoforos Anagnostopoulos
]]> 742 3 https://cdn.slidesharecdn.com/ss_thumbnails/mentatcanagnos-170810151953-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 https://public.slidesharecdn.com/v2/images/profile-picture.png My career has been motivated by a singular and genuine curiosity into what it means to learn from evidence, and in particular from evidence arising from measurements, i.e., data. This has driven me to study a variety of subjects, including mathematical logic, epistemology, probability theory, statistical modelling, machine learning and artificial intelligence. From a technical standpoint I am focused on the analysis of learning from streaming data in highly dynamic situations, ranging from cybersecurity to neuroimaging. Following an academic career as a Research Fellow at Cambridge University and an Associate Professor at Imperial College, I co-founded a startup in Machine Learning for da... http://www.canagnos.com https://cdn.slidesharecdn.com/ss_thumbnails/statisticalcyberisbis-180810071350-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/weakly-supervised-learning/109326652 Weakly supervised lear... https://cdn.slidesharecdn.com/ss_thumbnails/athensdatasciencemeetup-180810071012-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/why-data-science-is-a-science/109326276 Why Data Science is a ... https://cdn.slidesharecdn.com/ss_thumbnails/mentatcanagnos-170810151953-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/data-science-versus-artificial-intelligence-a-useful-distinction/78736791 Data Science versus Ar...