BPI cluster meeting – Participation behavior and social welfare in repeated task allocations

7th December 12:30-13:30 at Pav K.16 we have BPI cluster meeting.

Qing Chuan (Charlie) Ye from  Erasmus University Rotterdam will present his work on
Participation behavior and social welfare in repeated task allocations

Abstract:

Task allocation problems have focused on achieving one-shot optimality. In practice, many task allocation problems are of repeated nature, where the allocation outcome of previous rounds may influence the participation of agents in subsequent rounds, and consequently, the quality of the allocations in the long term. We investigate how allocation influences agents’ decision to participate using prospect theory, and simulate how agents’ participation affects the system’s long term social welfare. We compare two task allocation algorithms in this study, one only considering optimality in terms of costs and the other considering optimality in terms of primarily fairness and secondarily costs. The simulation results demonstrate that fairness incentivizes agents to keep participating and consequently leads to a higher social welfare.

Everyone is invited!

 

BPI cluster meetings!

BPI cluster meetings are back after the summer break.

Next meeting is on Wednesday September 7 at 12:30-13:30 in Pav. K16.

On Thursday, September 22 at 12:30-13:30 in Pav K.16 we have a special seminar!
Zuzanna Szymanska (ICM, University of Warsaw) will present her work about  “Computational modelling of cancer development and growth: Modelling at multiple scales  and multiscale modelling”.

Everyone is invited!

 

“Data Science from Different Angles” workshop

In the occasion of Julia Kiseleva’s PhD defense, there is a one day workshop “Data Science from Different Angles” at Eindhoven University of Technology on June 13. The invited speakers are Charles L.A. Clarke (University of Waterloo), Djoerd Hiemstra (University of Twente), Alexander Tuzhilin (NYU),  Wil van der Aalst (TU/e), Arjen P. De Vries (RUN) and Maarten de Rijke (UVA).

 

Please check the full program and register using the following link https://goo.gl/0dSAOR

BPi seminars are back!

BPI seminars will return after the winter break.

They will be held on the first Wednesday of the month, 12:30-13:00 in Pav K.16.

The next seminar is on May 4th, when Vasiliki Arvaniti will talk about “Data mining journal entries for discovering and analyzing unusual financial transactions”

We start also BPI discussion meetings.

They will be held on the third Wednesday of the month, 12:30-13:30 in Pav K.16.

Next meeting is on May 18th.

Everyone is welcome!

 

 

 

DSC/e Lecture Series

DSC/e Lecture Series: 11th November  (12:30-13:30) @ Grand Café ‘De Zwarte Doos’

Making your data “talk” to you:  linguistically conditioned models from data
by prof. dr. ir. Uzay Kaymak

Abstract:

With the abundance of data, many people concentrate on output-related properties of models, such as the accuracy. As long as the system, which processes the data, returns outcomes that are somehow acceptable (e.g. accurate), the underlying nature of the model that the system implements is considered secondary. Indeed, the system can learn from data, adapt itself to the data and hence “make the data talk”. However, natural language is an effective means for the end users to interact with the systems. Hence, there are advantages if our models from data can be conditioned on linguistic information. In this presentation, we consider models whose behavior can be understood in linguistic terms. The models typically return a set of linguistic rules or linguistic descriptions, which can be communicated to the user linguistically. This provides an additional means for the users to interact with the models learned from data.

Registration and more details: http://www.eventbrite.nl/e/dsce-lecture-series-november-11th-2015-uzay-kaymak-registration-16118399567

 

 

Next BPI cluster meeting – Friday 30th October

The BPI cluster meetings are taking place on the 5th Friday of the month, from 12:30 to 13:30.

Next meeting is on Friday, 30th October  12:30-13:30, Pav K16

Max van Rooijen will give presentation about evolving fuzzy system for printed circuit board (PCBA) demand forecasting.

Abstract. The author investigates the use of using an evolving fuzzy system for printed circuit board (PCBA) demand forecasting. The algorithm is based on the evolving Takagi–Sugeno (eTS) fuzzy system, which has the ability to incorporate new patterns by changing its internal structure in an on–line fashion. We argue that these capabilities could aid in forecasting dynamic demand patterns such as those experienced in the electronic manufacturing (EMS) industry. An eTS fuzzy system is implemented in the R statistical programming language and is tested on both synthetic and real–world data. To our knowledge, this is one of the first applications of an evolving fuzzy system to forecast product demand. The results indicate that the evolving fuzzy system outperforms competing approaches for the application considered.

 

Next BPI cluster meeting is on Friday 29th January.