Introduc*on!to!the!Course! -...
Transcript of Introduc*on!to!the!Course! -...
Machine Learning for Language Technology 2014
Introduc*on to the Course The Flipped Classroom
Marina San*ni [email protected]
Department of Linguis*cs and Philology Uppsala University, Uppsala, Sweden
Autumn 2014
Course Website & Contact Info: • hIp://stp.lingfil.uu.se/~san*nim/ml/2014/ml4lt_2014.htm
• Contact details: – [email protected] – [email protected] – [email protected]
Lecture 1: Introduction to the Course 2
Outline • Roll Call • Self-‐Presenta*on • Structure of the Course – People – About the Course – The Flipped Classroom – The Scalable Learning PlaUorm – Examina*on
• Learning Outcomes • Literature
Lecture 1: Introduction to the Course 3
Roll Call Candidate/Bachelor
• [email protected] • [email protected] • [email protected] • [email protected] • [email protected] • [email protected] • [email protected] • [email protected] • [email protected] • [email protected] • [email protected]
[email protected] • [email protected] • [email protected] • Avan.Sha-‐[email protected] • [email protected] • Emil.Wes*[email protected] • [email protected]
Master
• [email protected] • Qiuyue Quian ????????? • [email protected]
Lecture 1: Introduction to the Course 5
Professional Profile (LinkedIn)
• Computa*onal Linguist – Research scien*st (SICS East Swedish ICT) – Lecturer (Uppsala University)
• Research – Genre/Text Type Classifica*on
• WebGenreBlog – WebGenre R&D Group – etc. – Text Classifica*on/Sen*ment Analysis – Cross-‐Linguality – etc.
Lecture 1: Introduction to the Course 7
Bio
• I am Italian (from Rome) • got my PhD in the UK (Brighton University) • married to a Swede • living in Stockholm
Lecture 1: Introduction to the Course 8
COURSE STRUCTURE People, structure, design, mo*va*on, purpose
Lecture 1: Introduction to the Course 9
People
• Marina San*ni: responsible for the course as a whole, for the lab classes and assignments.
• • Joakim Nivre: decided the topics of the course and will deliver the online lectures.
• • Mats Dallhöf: responsible for all administra*ve issues related to this course.
Lecture 1: Introduction to the Course 10
About the Course
• Introduc*on to Machine Learning.
• Its focus is on methods used in Language Technology and NLP
• ML is a vast field… Selected topics
Lecture 1: Introduction to the Course 11
What is a ”flipped classroom”?
• Short answer: The flipped classroom inverts tradi*onal teaching methods, delivering instruc*on online outside of class and moving exercises into the classroom.
Lecture 1: Introduction to the Course 12
Flipping learning upside down • The basic idea is to reverse the structure of tradi*onal teaching.
• Tradi*onal teaching usually is based on: – lectures that are delivered in a classroom by a lecturer – homework carried out by students by themselves, not in the classroom.
• With the flipped approach, we will do the opposite: – you will listen to the online lectures at home, – you will be in the classroom to do your homework (that we will call lab sessions).
Lecture 1: Introduction to the Course 13
The Flipped Classroom Model • Students watch lectures at home at their own pace,
communica*ng with peers and teachers
• Concept engagement takes place in the classroom with the help of instructor.
• Basically, the flip teaching is a form of blended learning in which students learn new content online by watching video lectures, usually at home, and what used to be homework (assigned problems) is now done in class with teachers offering more personalized guidance and interac*on with students, instead of lecturing.
Lecture 1: Introduction to the Course 14
Learning Process
• Passive phase: that we can call the recep<ve phase, where the student/learner opens the mind by listening, reading and receiving new informa*on. In this phase the student lets new knowledge come in.
• Ac*ve phase: that we can call the produc<on phase, where the student/learner processes the new knowledge, constructs a personal concept map , creates cross-‐references with previous knowledge. In this phase, the student will become able to apply the new knowledge and to solve prac*cal tasks.
Lecture 1: Introduction to the Course 15
Research says that … … oren with tradi*onal teaching, where the passive phase is carried out in the classroom, learning outcomes are poor. For ex:
Lecture 1: Introduction to the Course 16
Thanks to Technology…
eLearning: thanks to the availability and success of online videos used for pedagogical purposes, and the increased access to technology, it is now possible to stop this nega*ve trend.
Lecture 1: Introduction to the Course 17
The Big Advantage • It allows students to personalize the learning at their own pace.
• You can replay the videos as many *me as you like, you stop them and resume them if you need to look up a word in a dic*onary, or if you need to brush up a concept, or if you are *red or hungry, etc.
• Therefore there is both a cogni*ve and physical advantage in doing the passive phase at home.
Lecture 1: Introduction to the Course 18
Success Story: Coursera
• The state of the art of online learning is MOOC, massive open online courses (wikipedia).
• Coursera (wikipedia) courses are a successful implementa*on of this idea: “more than one million people who have enrolled in the site’s courses are expected to pay aIen*on during video lectures interspersed with interac*ve exercises and complete homework assignments in between lectures” (source).
Lecture 1: Introduction to the Course 19
The Scalable Learning PlaUorm • In this course, we do not want to replicate or emulate Coursera.
• Our aim is to deliver a course (1) that requires a cogni*ve effort and (2) that can be learnt successfully both by candidate students and master students.
• With the flip teaching, we would like to allow students to adjust the learning of the new subject at their own pace and background knowledge.
• We will use plaUorm that has been developed in Sweden (by Swedish Ins*tute of Computer Science and Uppsala University) and it is called Scalable Learning.
Lecture 1: Introduction to the Course 20
Scalable Learning at Uppsala Uni
• The plaUorm is already successfully used at Uppsala University.
• David Black-‐Schaffer (Department of Informa*on Technology, UU) is regularly using it for his own courses.
• See David’s video presenta*on for mo*va*on, aims, and outcomes.
Lecture 1: Introduction to the Course 21
How are we going to work with the Scalable Learning plaUorm?
You need to: • create an account (only the first *me); • log in to the plaUorm when you receive a email sta*ng that a
lecture is ready. YOU WILL RECEIVE THIS EMAIL 2 OR 3 DAYS BEFORE THE RELATED LAB SESSION.
• Then:
– Home: listen to the video clips, answer the online quizzes, study related literature;
– Classroom: aIend the lab sessions and complete the lab tasks. This requires physical aIendance during the scheduled days (see schedule).
Lecture 1: Introduction to the Course 22
Analy*cs
• The plaUorm creates analy*cs that help the teacher to understand how the learning is going. These analy*cs are anonymous.
• The aim of this e-‐learning plaUorm is to understand which concepts and topics are more difficult for the students, thus enabling the teacher to provide the appropriate support.
Lecture 1: Introduction to the Course 24
Communica*on and Interac*on
• The plaUorm allows both anonymous and non-‐anonymous communica*on between students and teachers.
• The aim is to create an interac*on that is smooth, unproblema*c and seamless.
Lecture 1: Introduction to the Course 25
IMPORTANT! • If you do not aIend a video lecture on the plaUorm you
will not be able to carry out the tasks during the lab sessions. AIending the video lecture is a prerequisite to the related lab session.
• The comple*on of the lab tasks is compulsory.
• This is the basic structure of the course: 1. Home: Video Lecture + Quizzes + Reading 2. Classroom: Lab Tasks related to 1. Quizzes and Lab Tasks are not graded but must be completed in order to pass the course.
Lecture 1: Introduction to the Course 26
3 Graded Assignments • The course will be graded with three home assignments to be
completed individually and submiIed by the due date (see schedule).
• The idea is that you complete the assignments in the correct way but also with a certain degree of independent crea*vity. There are usually several different approaches to solve a problem, a task, or to complete an assignment. Choose the one that is more suitable for your mindset.
• Important always: – state and cri*cally discuss methodological assump*ons; – apply state-‐of-‐the-‐art methods we learn in this course; – present the results in a professionally adequate manner; use English
(scien*fic lingua franca) and academic style when wri*ng your reports.
Lecture 1: Introduction to the Course 27
Examina*on • Quizzes and Lab Tasks: The comple*on of quizzes and lab tasks is
mondatory. Quizzes and lab tasks are not graded.
• Assignments:The submission of each of the three home assignments is mondatory. Home assignments are graded and the following marks will be used: – Underkänd (U) [Fail] – Godkänt (G) [Pass] – Väl Godkänt (VG) [Dis*nc*on]
• In order to pass the course, three G are required + the comple*on of quizzes and lab tasks.
• In order to pass the course with dis*nc*on (VG), a student must pass at least two home assignments with dis*nc*on (VG)..
Lecture 1: Introduction to the Course 28
AIendance
• The whole aIendance requirement for the course is about 80%.
• This means that you should aIend 9 out of 12 online lectures and related lab sessions.
• If a student fails to fulfill this requirement, an addi*onal assignment will have to be completed prior to passing the course. The choice of the topic will relate to the missed material.
Lecture 1: Introduction to the Course 29
LEARNING OUTCOMES What the student will do that demonstrates learning
Lecture 1: Introduction to the Course 30
What is a learning outcome?
• Learning outcomes describe what students are able to demonstrate in terms of knowledge, skills, and values upon comple*on of a course.
Lecture 1: Introduction to the Course 31
Candidate/Bachelor
Arer the course, the student will be able to:
• apply basic machine learning principles to the linguis*c data; • apply methods to evaluate machine learning based systems performance within language technology;
• apply probability theory and principles of sta*s*cal inference to linguis*c data;
• use standard sorware for machine learning; • apply linear models for classifica*on; • apply clustering techniques to linguis*c data.
Lecture 1: Introduction to the Course 32
Master
Arer the course, the student will be able to:
• apply basic machine learning principles to the linguis*c data;
• apply probability theory and principles of sta*s*cal inference to linguis*c data;
• use standard sorware for machine learning; • implement linear models for classifica$on; • apply clustering techniques to linguis*c data.
Lecture 1: Introduction to the Course 33
Reading List (Required) • Text books
– Alpaydin E. (2010) – Daumé III H. 2012. – Jurafsky D. & Mar*n J. (2009) – Mitchell T. (1997). – Schay, Géza (2007)
• Papers – Androutsopoulos et al.(2000) – Collins (2002) – Metsis et al. (2006): only for master students – Nigam et al.(2000)
• For the Lab Sessions • Ian H. WiIen, Eibe Frank. 2005. • RapidMiner (maybe)
Lecture 1: Introduction to the Course 35