As ussual, the Google Calendar with all the Psst! dates is here:
http://www.google.com/calendar/embed?src=0tt13opsuigequfebur4e4dkac%40group.calendar.google.com&ctz=Europe/London
Monday, 24 October 2011
The beginning of Psst! seminars 2011-2012!
Dear CS Academics and PhD students
It is that time of the year when we start running our weekly Psst! seminars (PhD Students Short Talks).
Last year we had good (and constantly increasing) attendance. Seminars are usually fun and informative to attend, but remember that IT IS COMPULSORY TO ALL PHD STUDENTS TO GIVE ONE SUCH TALK EVERY YEAR.
FORMAT: as last year every speaker will have 15 minutes (including questions time), projector and laptop with ms PowerPoint will be provided. Seminars will happen in one of the lecture theatres. I will also bring tea and biscuits.
DATES: Below you can find the first version of the calendar (very likely to be changed). The most recent calendar is in the gDoc here: https://docs.google.com/spreadsheet/ccc?key=0Ano1xpvvLIiKdGpqbVlYbGlHaWtMOUU0b1UyX0Y0LWc&hl=en_US
WEEKDAY: last year we met on Thursdays 1:15PM, and if there will be no objections, we will continue meeting on Thursdays.
ORDER: as last year the 2nd year students will present first followed by everyone else. If you cannot present on a given date, please let me know as soon as you can.
++NOTICE TO ANYONE WHO WILL BE RUNNING STUDIES ANY TIME SOON:
++WE'VE BUILT A DIGITAL NOTICEBOARD TO ADVERTISE OUR STUDIES (bit.ly/noticeboard):
- over last few weeks we've build a digital noticeboard to advertise experiments to participants
- right now researchers can add studies, and students can sign up to email notifications.
- system is in beta testing, but we are planning to advertise it to students next week.
-if you are running a study, or about to run a study - please post it in the system
bit.ly/noticeboard
Pawel Orzechowski
CS PhD rep
bit.ly/noticeboard
bit.ly/pawelswebspace
It is that time of the year when we start running our weekly Psst! seminars (PhD Students Short Talks).
Last year we had good (and constantly increasing) attendance. Seminars are usually fun and informative to attend, but remember that IT IS COMPULSORY TO ALL PHD STUDENTS TO GIVE ONE SUCH TALK EVERY YEAR.
FORMAT: as last year every speaker will have 15 minutes (including questions time), projector and laptop with ms PowerPoint will be provided. Seminars will happen in one of the lecture theatres. I will also bring tea and biscuits.
DATES: Below you can find the first version of the calendar (very likely to be changed). The most recent calendar is in the gDoc here: https://docs.google.com/spreadsheet/ccc?key=0Ano1xpvvLIiKdGpqbVlYbGlHaWtMOUU0b1UyX0Y0LWc&hl=en_US
WEEKDAY: last year we met on Thursdays 1:15PM, and if there will be no objections, we will continue meeting on Thursdays.
ORDER: as last year the 2nd year students will present first followed by everyone else. If you cannot present on a given date, please let me know as soon as you can.
++NOTICE TO ANYONE WHO WILL BE RUNNING STUDIES ANY TIME SOON:
++WE'VE BUILT A DIGITAL NOTICEBOARD TO ADVERTISE OUR STUDIES (bit.ly/noticeboard):
- over last few weeks we've build a digital noticeboard to advertise experiments to participants
- right now researchers can add studies, and students can sign up to email notifications.
- system is in beta testing, but we are planning to advertise it to students next week.
-if you are running a study, or about to run a study - please post it in the system
bit.ly/noticeboard
Pawel Orzechowski
CS PhD rep
bit.ly/noticeboard
bit.ly/pawelswebspace
Thursday, 7 April 2011
Psst!5 Hamouda Chantar + Take a small part in organising seminars!
REWARDING JOB OPPORTUNITY (as I am leaving for few months)
this is a last seminar organised by me, at least for some time, as I am leaving for 2.5 months. I hope that seminars will continue uninterrupted, and that people of MACS will take some of my responsibilities in their own hands.
- There is a need for someone to take over sending emails
- and for someone to show up every time with a laptop and make sure that there is no chaos.
I am not sure if this could be maybe classified as lab helping, or other for an additional reward, but I guess you could try.
if you would be up for it, let me know soon
TODAY'S SPEAKER:
Hamouda Chantar - Document categorization
this is a last seminar organised by me, at least for some time, as I am leaving for 2.5 months. I hope that seminars will continue uninterrupted, and that people of MACS will take some of my responsibilities in their own hands.
- There is a need for someone to take over sending emails
- and for someone to show up every time with a laptop and make sure that there is no chaos.
I am not sure if this could be maybe classified as lab helping, or other for an additional reward, but I guess you could try.
if you would be up for it, let me know soon
TODAY'S SPEAKER:
Hamouda Chantar - Document categorization
-------------------------------------------------------------------
Abstract—Document categorization is an important
topic that is central to many applications that demand
reasoning about and organisation of text documents,
web pages, and so forth. Document classification is
commonly achieved by choosing appropriate features
(terms) and building a TFIDF document vector feature.
In this process, feature selection is a key factor in
determining the accuracy and effectiveness of resulting
classifications. For a given classification task, the right
choice of features means accurate classification with
suitable levels of computational efficiency. Meanwhile,
most document classification work is based on English
language documents. In this work (paper) we make three main
contributions: (i) we demonstrate successful document
classification in the context of Arabic documents
(although previous work has demonstrated text
classification in Arabic, the datasets used, and the
experimental setup, have not been revealed); (ii) we
offer our datasets to enable other researchers to
compare directly with our results; (iii) we demonstrate
a combination of Binary Particle Swarm Optimization
and K nearest neighbour that performs well in selecting
good sets of features for this task.
topic that is central to many applications that demand
reasoning about and organisation of text documents,
web pages, and so forth. Document classification is
commonly achieved by choosing appropriate features
(terms) and building a TFIDF document vector feature.
In this process, feature selection is a key factor in
determining the accuracy and effectiveness of resulting
classifications. For a given classification task, the right
choice of features means accurate classification with
suitable levels of computational efficiency. Meanwhile,
most document classification work is based on English
language documents. In this work (paper) we make three main
contributions: (i) we demonstrate successful document
classification in the context of Arabic documents
(although previous work has demonstrated text
classification in Arabic, the datasets used, and the
experimental setup, have not been revealed); (ii) we
offer our datasets to enable other researchers to
compare directly with our results; (iii) we demonstrate
a combination of Binary Particle Swarm Optimization
and K nearest neighbour that performs well in selecting
good sets of features for this task.
Monday, 28 March 2011
Psst!4 - Thursday 31st April, 1:15 G.44
Govan, Gordon : Constructing Metabolic Networks from Mass Spectrometry Analysis
Metabolic networks are often constructed through a long process that involves inspecting the genetic material of the cell being analysed to find genes which correspond to metabolites.
Mass spectrometry analysis quickly separates and measures the masses of the chemicals within a cell. From this data we can construct a network and then we try to validate this model using graph analysis techniques.
Ritter, Christopher: Autonomous emotional agents
(TBA)
- building autonomous emotional agents (based on the FATIMA agent architecture)utilising traditional planning and emotional appraisal for better dialogues; including turn taking and more complex conversation models allowing for modelling simple opinion making/defence processes
- examine the utilisation of such agents in the Serious Games Approach (Socio-Critical/-Educational Games)
i am looking forward to seeing you all there,
as ussual tea, coffee and bisquits will be served
Pawel Orzechowski
MACS CS PhD rep
Metabolic networks are often constructed through a long process that involves inspecting the genetic material of the cell being analysed to find genes which correspond to metabolites.
Mass spectrometry analysis quickly separates and measures the masses of the chemicals within a cell. From this data we can construct a network and then we try to validate this model using graph analysis techniques.
Ritter, Christopher: Autonomous emotional agents
(TBA)
- building autonomous emotional agents (based on the FATIMA agent architecture)utilising traditional planning and emotional appraisal for better dialogues; including turn taking and more complex conversation models allowing for modelling simple opinion making/defence processes
- examine the utilisation of such agents in the Serious Games Approach (Socio-Critical/-Educational Games)
i am looking forward to seeing you all there,
as ussual tea, coffee and bisquits will be served
Pawel Orzechowski
MACS CS PhD rep
Monday, 21 March 2011
Event: Monday March 28th , 1:15, room 3.06 Psst! 3 PhD students' short talks 3: Marta Vallejo, Robert Stewart
Marta Vallejo: Computational approaches to Uncertainty in Scenario-based models of Climate Change Impacts
My PhD will try to fill the gap between local and global modelling approach by scaling up part of the current ecological models to a global perspective. From this global point of departure, the main focus of the project is the study of how policies, regulations and guidelines can modify the behaviour of the stakeholders involved in the use of landscape and subsequently infering the effects in the physical dynamics of the land cover.
Robert Stewart: Fault Tolerant Symbolic Computation in Exascale Distributed Systems
The past decade has seen distributed computing move into the realms of exascale, heterogeneous networked computer systems. This evolutionary progression has redefined the problems faced for parallel execution on dependable computational systems.
My PhD research is focused on adding fault tolerance at the software level, to satisfy the dependable requirements on such platforms. The ambition is that the outcomes of my PhD will contribute to the HPC-GAP project (High Performance Computational Algebra and Discrete Mathematics), providing a set of fault tolerant mechanisms to overcome hardware failure, byzantine failures, and unreliable heterogeneous topologies. There are a number of target platforms available for developing these mechanisms, including the HECToR cluster in Edinburgh: "The UK National Supercomputing Service", that will ultimately scale up to a million processor cores.
My PhD will try to fill the gap between local and global modelling approach by scaling up part of the current ecological models to a global perspective. From this global point of departure, the main focus of the project is the study of how policies, regulations and guidelines can modify the behaviour of the stakeholders involved in the use of landscape and subsequently infering the effects in the physical dynamics of the land cover.
Robert Stewart: Fault Tolerant Symbolic Computation in Exascale Distributed Systems
The past decade has seen distributed computing move into the realms of exascale, heterogeneous networked computer systems. This evolutionary progression has redefined the problems faced for parallel execution on dependable computational systems.
My PhD research is focused on adding fault tolerance at the software level, to satisfy the dependable requirements on such platforms. The ambition is that the outcomes of my PhD will contribute to the HPC-GAP project (High Performance Computational Algebra and Discrete Mathematics), providing a set of fault tolerant mechanisms to overcome hardware failure, byzantine failures, and unreliable heterogeneous topologies. There are a number of target platforms available for developing these mechanisms, including the HECToR cluster in Edinburgh: "The UK National Supercomputing Service", that will ultimately scale up to a million processor cores.
Monday, 7 March 2011
Event: Thursday March 10rd , 1:15, room G.44, Psst! 2 PhD students' short talks 2: Michael Kriegel, Andrew MacVeen, Allan Weallans
Michael Kriegel: Learning From Stories
The goal of computer scientists working on Interactive Storytelling is to create computer programs that can improvise stories in real time taking into account input from the audience. While many algorithms have been suggested and systems were designed, almost none of them has managed to convincingly improvise stories in practise. This can be in large parts attributed to the difficulty of capturing the knowledge about the story world that these algorithms need to perform. In this talk I am going to give an overview of my PhD work in which I am trying to solve this knowledge acquisition problem by automatically extracting a large portion of the required knowledge from a set of example stories set within the story world.
Allan Weallans: Distributed Drama Management: Dramatic Intelligence in Emergent Narrative
Abstract: Emergent Narrative is a character-based approach to interactive digital storytelling with the aim of diminishing the effect of the Narrative Paradox - the inversely proportional relationship between authorial control of the story and user freedom within the story. However, Emergent Narrative in its current state is highly simulative: characters will select courses of action that are believable within the context of the story and the character's own personality, but they are not guaranteed to select dramatically interesting actions, much less sequences of dramatically interesting actions occurring in a recognisable dramatic structure. Our work, DISTRIBUTED DRAMA MANAGEMENT, intends to build on existing Emergent Narrative work by incorporating "dramatic intelligence" into our character agents by monitoring the story's dramatic needs and giving agents the ability to co-ordinate and re-prioritise their characters' goals for dramatic purposes. In this way, we can retain the character believability and action contextualisation that are the strengths of Emergent Narrative while allowing the characters to be steered towards fulfilling given narrative concerns such as dramatic tension, structure and theme.
The goal of computer scientists working on Interactive Storytelling is to create computer programs that can improvise stories in real time taking into account input from the audience. While many algorithms have been suggested and systems were designed, almost none of them has managed to convincingly improvise stories in practise. This can be in large parts attributed to the difficulty of capturing the knowledge about the story world that these algorithms need to perform. In this talk I am going to give an overview of my PhD work in which I am trying to solve this knowledge acquisition problem by automatically extracting a large portion of the required knowledge from a set of example stories set within the story world.
Andrew MacVeen: Task-Involved Versus Ego-Involved: Motivating Children to Exercise In a Pervasive Exergame
This talk presents current work on a pervasive health and fitness game (exergame), designed to motivate children to reach their recommended daily exercise goals and facilitate long term behavioural change. I discuss briefly the current work in the area of pervasive exergames and highlight a common theme in the approach they take. Through my study of the relevant psychology literature, I identify a potential problem in this approach - the goal context that the majority of systems adopt. I hypothesize that a 'Task-Involved' rather than an 'Ego- Involved' system would be more suitable at addressing the problem of sedentary childhood behaviour.
Allan Weallans: Distributed Drama Management: Dramatic Intelligence in Emergent Narrative
Abstract: Emergent Narrative is a character-based approach to interactive digital storytelling with the aim of diminishing the effect of the Narrative Paradox - the inversely proportional relationship between authorial control of the story and user freedom within the story. However, Emergent Narrative in its current state is highly simulative: characters will select courses of action that are believable within the context of the story and the character's own personality, but they are not guaranteed to select dramatically interesting actions, much less sequences of dramatically interesting actions occurring in a recognisable dramatic structure. Our work, DISTRIBUTED DRAMA MANAGEMENT, intends to build on existing Emergent Narrative work by incorporating "dramatic intelligence" into our character agents by monitoring the story's dramatic needs and giving agents the ability to co-ordinate and re-prioritise their characters' goals for dramatic purposes. In this way, we can retain the character believability and action contextualisation that are the strengths of Emergent Narrative while allowing the characters to be steered towards fulfilling given narrative concerns such as dramatic tension, structure and theme.
Wednesday, 2 March 2011
March 3rd,1:15, G.44: Psst! PhD students' short talks 1: Thomas Methven & Pawel Orzechowski
Psst! 1 will happen on March 3rd, 1:15, G.44
Thomas Methven: Does 3D Affect Gloss Perception?
When we look at a real surface, we see two types of disparity which
help us define what we're looking at. Both eyes get slightly different
views on the geometry, helping us see it in 3D. In addition, both eyes
see a slightly different pattern of gloss. How important are these two
types of disparity to the realism of the surface and how can we
manipulate them? In this talk, I will briefly cover my current PhD
work on this topic.
Pawel Orzechowski : Interactive mobile presentation of textiles.
While buying textile products using a handheld device, buyers face the perceptual gap between qualities they can perceive via the interface and those sensed when handling the real product. In this research I investigate the textile qualities naive shoppers look for and the gestures they use to handle fabrics. Finally, I use movement patterns to prototype mobile touch-screen interactive-video interfaces and communicate the desired qualities better.
Thomas Methven: Does 3D Affect Gloss Perception?
When we look at a real surface, we see two types of disparity which
help us define what we're looking at. Both eyes get slightly different
views on the geometry, helping us see it in 3D. In addition, both eyes
see a slightly different pattern of gloss. How important are these two
types of disparity to the realism of the surface and how can we
manipulate them? In this talk, I will briefly cover my current PhD
work on this topic.
Pawel Orzechowski : Interactive mobile presentation of textiles.
While buying textile products using a handheld device, buyers face the perceptual gap between qualities they can perceive via the interface and those sensed when handling the real product. In this research I investigate the textile qualities naive shoppers look for and the gestures they use to handle fabrics. Finally, I use movement patterns to prototype mobile touch-screen interactive-video interfaces and communicate the desired qualities better.
Subscribe to:
Posts (Atom)