BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//euroscipy-2026//speaker//MR8NT3
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-euroscipy-2026-KFBJXK@pretalx.com
DTSTART;TZID=CET:20260720T152000
DTEND;TZID=CET:20260720T155000
DESCRIPTION:Mapping freeform research affiliations to persistent identifier
 s such as [ROR (Research Organization Registry)](https://ror.org/)  is har
 der than it looks. Institution names appear in many forms such as abbrevia
 tions\, alternate spellings\, local languages\, or legacy names\, thus mak
 ing a reliable mapping difficult to achieve at scale.\n\nIn this talk\, we
  present a semantic retrieval pipeline that reframes institution identific
 ation as a search problem rather than a string-matching task. Our system c
 ombines named entity recognition to extract institution entities\, dense e
 mbeddings to represent their semantic meaning\, and vector search to retri
 eve the most likely ROR matches. This approach allows us to handle noisy\,
  incomplete\, and multilingual inputs while remaining resilient to variati
 on in how institutions are referenced.\n\nBy treating institution matching
  as semantic retrieval\, we improve recall and robustness without relying 
 on heuristics or on a continuous expanding rule-based approach. The system
  scales naturally as new institutions are added and as naming conventions 
 evolve\, making it well suited for the dynamic research environment.\n\nWe
  will share implementation details\, evaluation results\, and practical le
 ssons learned from deploying this pipeline in a real-world production sett
 ing.
DTSTAMP:20260603T191744Z
LOCATION:Room 1.19 (Ground Floor\, Shannon)
SUMMARY:Finding the Right ROR: Semantic Search for Research Institutions - 
 Diogo Rodrigues\, Frank Sauerburger
URL:https://pretalx.com/euroscipy-2026/talk/KFBJXK/
END:VEVENT
BEGIN:VEVENT
UID:pretalx-euroscipy-2026-RBUGDR@pretalx.com
DTSTART;TZID=CET:20260721T152000
DTEND;TZID=CET:20260721T154000
DESCRIPTION:Recent advances in AI have shifted industries’ attention towa
 rd integrating LLM-based systems. Even though LLMs can solve a wide range 
 of business problems\, they came with a significant complexity overhead. A
 t same time\, many real-world business applications involve well-defined o
 bjectives\, predictable inputs\, and clear evaluation criteria. \n\nToday\
 , we are increasingly seeing a default pattern: for almost any NLP use cas
 e\, teams prompt GPT-like models and pay the bill at the end of the month.
  However\, this approach often introduces unnecessary complexity\, costs\,
  and operational risk. Many business and research problems exist in constr
 ained environments that can be solved with simpler techniques\, achieving 
 the same or higher success rates.\n\nThis talk defends that fine-tuned BER
 T-based models remain a strong and often superior choice for targeted busi
 ness use cases that require NLP-based solutions. I propose to present a re
 al\, in-production use case where a simple transformer-based classifier de
 monstrates a more favourable performance-cost trade-off than LLM-based app
 roaches\, driven by lower latency\, reduced operational complexity\, easie
 r fine-tuning\, and significantly lower maintenance costs.\n\nThe goal of 
 this presentation is not to reject LLMs\, but to promote a pragmatic\, out
 come-driven approach to NLP\, where “boring” solutions often deliver t
 he most value.
DTSTAMP:20260603T191744Z
LOCATION:Room 2.41 (First Floor\, Turing)
SUMMARY:Boring AI Works: When BERT Beats Billion-Parameter Models - Diogo R
 odrigues
URL:https://pretalx.com/euroscipy-2026/talk/RBUGDR/
END:VEVENT
END:VCALENDAR