24.02.2010 The 'Versandhausberater' presents the second convention for IT in mail order business

Mail Order, reloaded: IT-concepts for tomorrows mail-order

This is the title of the second convention on IT in the mail-order business which takes place on Feb, 23rd + 24th 2010 in Frankfurt.

Among other top-class contributors Prof. Dr. Feindt will give a talk on "Moderne Prognosetechnologien für den Versandhandel" ("modern prediction technologies for mail-order business"). Beside the convention there is a highly specialised fair, where operator of onlineshops, retailers and manager of german mail-order companies can catch up on mail-order systems, software and CRM-solutions.

Further information can be found at www.it-at-commerce.de, where you can also register yourself.

24.12.2009 The FAZ reports on the NeuroBayes Short Term Trading Fund

On Dec 23rd the Frankfurter Allgemeine Zeitung reported on the NeuroBayes Short Term Trading Fund.

The NeuroBayes Short Term Trading Fund was intiated by Lupus Alpha three years ago in a talent-hotel, as described in the article. Since then it is managed by <Phi-T>. The tradings strategies were developed by professor Michael Feindt. They are based on techniques which are used in the fields of high energy physics. The success proves these strategies to be right.

31.08.2009 Lebensmittelzeitung reports on NeuroBayes® at Otto mail order business

In their release 34/09 the Lebensmittelzeitung reports on enormous savings the Otto group made by using NeuroBayes^®.

With very preciese predictions from NeuroBayes^® on the expected sales for every single article that Otto, Baur and Sportscheck offer, the concern will save about 20 Mio Euro in the running business year. This is possible because the stock overrun can be minimized when the expected sales rates for each article are predicted more precisely. Also more articles can be sold because they are more likely to be in stock when ordered by the customer.

You can find the whole article in german below this text.

31.08.2009 Hurry away from competitors with artificial intelligence

In the release from september 2009 OnetoOne reports on how the Otto group wants to increase multi channel sales with a joint venture with Phi-T.

You can find the article in german below this text.

28.08.2009 Otto models multichannel behaviour

The business journal Versandhausberater reports in the release from august 28th 2009, how the joint venture of Phi-T and Otto, Phi-T products & services, can predict sales numbers more precisly and can help to optimise the advertising mixture.

Den Artikel finden sie unterhalb dieses Textes

27.08.2009 Shopping schedule with elemenatry particles

In the release from august 27th 2009 the business journal Textilwirtschaft reports on the success the Otto group retrieved by using NeuroBayes® for the sales predicion.

After the usage of NeuroBayes® predictions for mailorder business of Otto has proven its power, there are plans to offer this technology to other companies.

You can find a copy of this article in german below this text.

26.08.2009 Phi-T sponsors the gridka summer school 2009

During the week between the 31st of august and the 4th of semptembre 2009 there will be once again the gridka school, which is again sponsered by <Phi-T>.

At the gridka school administrators and experienced users of the GRID will meet at campus north of the Karlsruhe Institut for Technology and get to know new tips and tools around the grid computing in talks and tutorials.

The GRID is a network of computing centres around the world which offer their computing power to scientists from different fields. Main user and provider of the GRID is CERN and the associated institues, because the enormous amounts of data the LHC will produce can only be processed in the GRID.

Because NeuroBayes^®  will be one of the tools used for data evaluation and will be used in the grid, <Phi-T> supports the gridka school as one of the main sponsors (beside Microsoft Research and Sun). Thus there will be social events in the evening where the paricipants can exchange their experiences.

Prof. Dr. Feindt will give a talk at the summer school on method transfer from science to business.

<Phi-T> wishes all participants a good time and interesting conversations at the gridka school 2009.

26.08.2009 Phi-T sponsors the flavianet summerschool in Karlsruhe

The summer school on flavour physics taking place from 07th till 18th septembre 2009 at the Karlsruhe Instutit für Technologie is sponsored by <Phi-T>.

Flavour physics is dealing with the basic modules of mater, the so called "quarks", and the forces between them. At the flavianet summerschool young scientists from all around the world will meet in order to exchange their knowledge on theoretical predictions as well as on experimental measurements of bound states of quarks and their special properties.

It is <Phi-T>'s special concern to support this summerschool, because the basics of the multivariate analysis tool NeuroBayes^® were developed during the research of Prof. Dr. Michael Feindt on flavour physics, precisly on b-physics.

Prof. Dr. Feindt will give a talk entitled "Artificial intelligence for flavour physics and economy" at the summer school which is about method transfer from science to business.

<Phi-T> wishes all participants to have a good time and have intersting discussions at the flavianet school 2009.

25.08.2009 Spiegel reports about the Phi-T Lupus Alpha fund

DER SPIEGEL reports in his release (35/2009) about "algotrading" in an article called 'Alpha-Physiker an der Börse' (Alpha-physicists at the stock exchange).

Among others, Prof. Michael Feindt was interviewed for this article, because with his fund managed by Phi-T he is one of the leaders in the field of purely algorithmis trading. 

The subtitle of the article is asking:

"Computers take over the work of stock traders. Within milliseconds they fire their orders to the stock exchanges. Can this destabilise the markets?"

You can read the full article here (german) and find the answer to this question.
 

21.08.2009 Particle physicists make trading more efficent

In the release from aug 21st, 2009, the Handelsblatt reports about how the Otto group is optimising their warehouse stock using neural networks
 

The primary project for the optimisation of the mail order business at Otto lead to an enormous cost reduction. It was so successful, that Otto launched together with Phi-T a common joint venture, Phi-T P&S, in order to ensure this know how in the long run for their business.
 

25.06.2009 DataMiningCup

Universität Karlsruhe (TH) wins with assistence of NeuroBayes®

This year once again a team of physics students of the University of Karlsruhe participated in the Data Mining Cup 2009 and won the first price.

The student team with its driving forces Michael Prim, Fabian Keller and Johannes Munk was made up of participants of the lecture "Moderne Methoden der Datenanalyse" ("modern methods of data analysis“) given by the elementary particle physicist Prof. Dr. Michael Feindt. In the past four years students supervised by Prof. Dr. Feindt have regularly succeeded in taking leading positions in the Data Mining Cup, e.g. in 2006 ranks 1 - 4, when the prices achieved in ebay auctions had to be predicted.  

Prof. Dr. Feindt is the inventor of NeuroBayes®, the forecast-algorithm which we continually develop and optimise for commercial applications at Phi-T. This year the program was once again put at the students disposal. 

The anual Data Mining Cup was first held ten years ago and is the worlds largest students competition in the field of intelligent data analysis and forecasting. This year however, the prudsys AG, that hosts the Data Mining Cup, was for the first time not looking for the best individual student data miner. Instead, the universities themselves competed against each other in teams. The success was large: More than 100 universities from home and abroad participated, including universities from Canada, USA, UK, Russia and China. The competition ended on may, 25^th 2009 with 52 solutions handed in.

As every year a corporate user supported the Data Mining Cup and provided the task definitions as well as the data. This years task came from the field of books wholesale where the Libri GmbH presented the challenge of predicting the sales volume of 8 items in roughly 2,400 shops.

In the dataset for training the sales volumes of 1,856 product groups as well as the 8 target items were listed for about 2,400 stores. Based on the knowledge of sales of the product groups the sales for the 8 target items were supposed to be predicted for approximately 2,400 further stores. One of the main difficulties was the huge variance in the considered orders of magnitude. This was true not only for the shops - their sizes ranging from huge e-tailers down to tiny corner shops - but also for the size of the product groups.

You can find more information concerning the Data Mining Cup under www.data-mining-cup.de.

 

08.06.2009 Public lecture series at the University of Karlsruhe (KIT)

Journey to the Limtits of Knowledge: Smallest Particles - highest Energies - highest Precision. 

This is the title of a block of public lecture for laymen in the traditional ''Saturday Phyiscs'' series at the University of Karlsruhe. Experts explain recent results and research projects in elementary particle physics in an understandable way. Also Phi-T founder Prof. Dr. Michael Feindt will give a presentation. Don#T miss, it will be interesting... 

 

 

Programme:

27. Juni 2009	Prof. Dr. Ulrich Nierste Vom Weltall zur Weltformel: Rätsel der modernen Physik
04. Juli 2009	Prof. Dr. Michael Feindt Gespiegelte Antiwelten: Die experimentelle Bestätigung der Nobelpreis-2008-Theorie
11. Juli 2009	Prof. Dr. Guido Drexlin Boten aus dem Universum – Neutrinos, dunkle Materie und kosmische Strahlung
18. Juli 2009	Prof. Dr. Thomas Müller Der Large-Hadron-Collider am CERN: dem Urknall auf der Spur

10.00 - 11.30 o'clock, Gaede-Hörsaal of the University of Karlsruhe, free entrance.

Flyer: www.physik.uni-karlsruhe.de/Aktuelles/Physik_am_Samstag/PhysSams09.pdf

Web:  www.physik.uni-karlsruhe.de/Aktuelles/Physik_am_Samstag/

 

 

04.06.2009 Investment Conference London, June 12, 2009

Phi-T founder Prof. Dr. Michael Feindt will give the keynote address with the title  Extremely Quantitative: Future Predictions for CERN, Soccer, Fashion and Financial Markets at this year's  Battle of the Quants in London: At the conference leading minds of the investment industry are discussing ways of how to use quantitative methods to overcome the financial crisis.

www.battleofthequants.com

27.05.2009 Phi-T mousegame

 Find out was a mouseclick has in common with a stock price und see how predictable your free will is.

Like the most humans you might think too, that a computer is unable to predict which key you are going to press next. Find out here, that it might be possible and what all this has in common with stock prices.

30.04.2009 LebensmittelZeitung

In its edition dated April 17, 2009 the weekly paper LebensmittelZeitung published the article "Künstliche Intelligenz arbeitet für dm" and reported about the use of NeuroBayes^® at the chain dm-Drogeriemarkt as well as about the Joint Venture of Phi-T and Otto Group.

This link leads you to the original article of Jörg Rode.

22.04.2009 Neuro Netzer Bundesliga

The NeuroBayes^® "Neuro Netzer" predicts the results of the first german soccer league "Bundesliga".

NeuroBayes^® is not only suitable for "serious" science, but also for more funny things.

On basis of historical data a program has been developed, which predicts the outcome of games of the first german soccer league. It calculates the probality for each team to win or if it will be drawn.

In the latest forecast it tells that the VfL Wolfsburg hast an over 50% probablity to win against Engergie Cottbus. We'll see if this is correct...

24.03.2009 Karlsruhe

Fermilab experiments constrain Higgs mass. Important parts of the measurement profit from NeuroBayes^®.
 

For particle physicists it is one if the last puzzle pieces missing to complete the theory known to them as the “Standard Model”. This Standard Model describes the structure of all matter, the properties of the particles making up matter and their interactions. Solely the origin of the mass of the elementary particles (fermions and bosons) cannot yet be explained adequately by the Standard Model. A possible explanation is given by the so called “Higgs Mechanism”. This claims that the masses of particles originate from their interaction with the Higgs Field. For years now, particle physicists from all over the globe have been trying to find the Higgs particle. Up until now though, it has proved elusive.

Now, scientists at the Fermlab (Chicago) have succeeded in constraining the possible mass range for the Higgs particle -should it actually exist- even further. This was achieved by combining many individual measurements at the experiments CDF and DZero.

For some of the particularly tricky measurements NeuroBayes^® played the decisive role and thus contributed greatly to the quality and precision of the final result.

Further measurements at the particle accelerator LHC at CERN, that is anticipated to start up this autumn, are supposed to answer the question of the existence of the Higgs particle in the coming years. Also in these future measurements NeuroBayes^® will most likely make a significant contribution.

Press release from Fermilab

 

09.03.2009 Karlsruhe

NeuroBayes ^® makes an important contribution to the long sought for discovery of rare process.

According to the current physical theory of the world, matter is made up of six different quarks and six leptons. Only the lightest of these elementary particles make up the world surrounding us, whereas the heavier ones decayed shortly after the big bang. Evidence for these particles can only be found at powerful particle accelerators.

Physical theories are always then put to a hard test when expected processes prove very hard to find. The production of single top quarks -the heaviest known quark type- was such an eagerly sought for discovery. This process is so rare that it is masked by others that are a lot more frequent. Therefore to find it, it was not sufficient to collect further data from collisions at the particle accelerator of the Fermilab. A complex system of ingenious implements for identifying this rare process also had to be developed.

In this system NeuroBayes^® is employed simultaneously at several points and was thus able to contribute essentially the discovery, with which physicists once again succeeded in confirming their expectations and thus our established view on the world.

You can find the CDF-website under this link link and the publication here.

 

 

11.12.2008 Karlsruhe

Fermilab Result of the Week: NeuroBayes^® finds evidence for new matter.

In modern physics two fundamental types of matter are known. On the one hand there are “baryons” like for instance the proton and the neutron, on the other hand there are “mesons” such as the pion. These two types of matter differ by the number of quarks of which they consist - baryons are made up of three, mesons of only two quarks.

Furthermore it is known that barons can form bound states with each other. Protons and neutrons agglomerate to create the elements of the periodic table.

The mysterious X(3872)-particle is an enthralling topic of current research. It was first seen at the Belle-experiment in Japan and has since been sighted at the CDF-experiment in the USA. Its properties cannot be explained by the hitherto known forms of matter. On the other hand these properties could be explained perfectly if X(3872) was in truth a new sort of matter, namely a bound state of two mesons.

Physicists at the university of Karlsruhe participating in the CDF-experiment have analysed this peculiar particle using NeuroBayes^®. Therby the world best measurement of its mass was achieved which supports the theory of new matter.

 You can find more information on this exciting topic under he"Who is Mister X?" auf http://www.fnal.gov/pub/today/archive_2008/today08-12-04.html.