TasaractDavut Emre Taşar
Naijá

My profile

I be data scientist for BBVA AI Innovation team for Madrid. For my PhD for University of Navarra, I dey study whether banks fit make fairer decisions if dem fit measure how sure their AI predictions be.

See di raw data JSON

Small tori about me

I come from İzmir, for Turkey side of di Aegean Sea. I bin be civil engineer till one arrow hit me for knee. I swap my hard hat for lab coat. Now I be data scientist for BBVA AI Innovation team for Madrid.

How much trust bank fit put for wetin AI model predict? For my PhD for University of Navarra, I dey study whether measuring and reducing dat uncertainty fit help banks make fairer decisions. When I study strategic management for Oxford, I dey wonder whether today’s decision go still work if things change tomorrow.

For abstract maths, different problems fit get di same pattern. Sometimes I dey read quantum physics, start to wonder wetin a measurement really tell us, and before I know am I don enter philosophy. I like how ideas connect like dat.

I still dey explore Madrid. As my Spanish dey improve, I dey learn more about Spain and Latin American cultures. I like to paint and play chess too.

For house I dey make coffee well-well. I get small darbuka drum, kalimba and one Zanpakuto from Bleach wey I dey use for kendo practice. If I get chance, I dey make up stories to make my daughter laugh.

Oxford Strategic Management Executive Programme

Where I dey work and study

BBVA · AI Transformation · AI Innovation
Engineer and scientist for applied research
University of Navarra
PhD researcher wey dey study how AI fit treat people fairly

Wetin interest me

Ontologies and data structures

Wetin go change for graph if one train come late?

Station na place. Train departure na event. If I use different node type for each one and keep where each time come from, I fit trace all the connections wey one delay go affect?

Imagine say two trains suppose leave the same station by nine. For graph, I fit put the station there once and give each train departure im own record. That way, if one train time change, the other one time go still dey as e be.

I go keep the timetable time, the new time and where each one come from. If one departure move from nine to ten minutes past nine, I fit follow the links from that departure to see which connecting trains I need check. If I write only one time for the whole station, that difference go lost.

RDF 1.1 · W3C
Meaning-Native Computing

If computer know the rules wey one program must follow, e fit predict better which instructions go run next?

Meaning-Native Computing na my own independent research project. I wan compare two ways to predict wetin go happen, with the same program and about the same computing power. I go give one of dem the program rules too, write am out clear. E go help am see the next steps better? Before any change wey e predict fit take effect, another check must confirm say the change follow the rules.

For this research wey I dey do on my own, I wan compare two ways to predict wetin one program go do. Both go watch the same program as e dey run, with about the same computing power. One go get the program rules too, as ontology wey computer fit run. If e know the rules, e go fit tell better which instructions go run next?

The rules fit allow two different things to happen next. Person fit predict the wrong one even when the prediction follow every rule. I wan check the prediction against wetin the program really do. Before any change wey the method suggest take effect, another check wey no depend on the prediction must show say the rules allow am.

Meaning-Native Computing
Category theory

Wetin go remain when I join two transformations?

One process turn timetable to graph. Another one find route for the graph. When I join the two, how I fit keep track of which time belong to which station?

One process turn train timetable into graph. Another one find route through the graph. When I join the two so the second use wetin the first produce, na example of composition. I wan know whether I still fit trace every time wey dey the final route back to the correct station.

Suppose one train dey leave by nine for each of two stations. If the first process keep only the time and throway the station, the process wey dey find route no fit use “nine” alone to recover the link. Even small timetable fit make me think about wetin each step need pass to the next one.

Seven Sketches in Compositionality
Quantum measurement

Wetin go happen if I ask the same qubit another question?

I prepare the “+” state. If I measure for X basis, na + I go always get. For Z basis, 0 and 1 get equal chance. Wetin dey fascinate me be say the answer dey change with how I measure am, even when I prepare am the same way.

Imagine say before every measurement I prepare one qubit again for the plus state. If I measure am for X basis under ideal conditions, I go get plus every time. If I prepare am the same way and measure am for Z basis, 0 and 1 get equal chance to show.

If I measure ten times for Z basis, I fit get seven zeros and three ones. Zero get half chance each time. E no mean say exactly five zeros must show for ten tries. E easy to miss that difference if na just small set of results you dey look.

Quantum Information · IBM · Specify observables in the Pauli basis · IBM Quantum Documentation

Award for my work

For 2024, BBVA give me the We Think Big award for my work on patents and wetin I contribute to academic research.

See the award photo

How to reach me and see my profiles