@MadelonHulsebos Profile picture

Madelon Hulsebos 🦋 madelonhulsebos.bsky.social

@MadelonHulsebos

Researching neural models for structured data; table representation learning💫; faculty @cwinl member @ellisforeurope, prev @berkeley_eecs @uva_amsterdam

Similar User
Dutch Seminar on Data Systems Design photo

@dsdsdnl

Stefan Grafberger photo

@SGrafberger

Andreas Kipf photo

@andreaskipf

Anna Fariha photo

@FarihaAnna

Peter Boncz photo

@peterabcz

SIGMOD/PODS 2025 photo

@SIGMODConf

Table Representation Learning @NeurIPS photo

@TrlWorkshop

Sebastian photo

@sscdotopen

Calc Consulting photo

@CalcCon

Immanuel Trummer photo

@ImmanuelTrummer

Carsten Binnig photo

@cbinnig

CIDR 2025 photo

@cidrdb

Matthias Boehm photo

@matthiasboehm7

Intelligent Data Engineering Lab photo

@INDE_LAB_AMS

Pınar Tözün (@pinar@social.itu.dk) photo

@pinartozun

Pinned

Tables are a goldmine for accurate, fresh, domain data that LLMs should be grounded in for RAG, factver/QA, text2sql: retrieval is key! We introduce 🎯TARGET, a benchmark for evaluating table retrieval! Code,data,paper: target-benchmark.github.io Thinking bm25 ftw? Think twice..🧵

Tweet Image 1

Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

I'm now also on bluesky 🦋 bsky.app/profile/madelo…. Will just duplicate for now, but feels pretty promising!


Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

- Read more in the paper: arxiv.org/abs/2406.19380… - Use TabReD in your new method evals: github.com/yandex-researc… - Follow other people who are doing great work in this direction: @__smarton @__mfeurer__ @jpgard @MadelonHulsebos @TrlWorkshop <|eos|>


Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

Tabular DL success on benchmarks ≠ success in production. We know this first hand, trying to ship models. This motivated us to create TabReD - a new suite of 8 tabular datasets that capture real-world data characteristics overlooked by existing benchmarks. (1/N)

Tweet Image 1

Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

As @wenhuchen once put it: “...if your retrieval is mediocre, your LLM can easily backfire”. We should not overlook the value of structured data! Let's work on better retrieval for structured data; contrib and eval new table retrievers w 🎯TARGET target-benchmark.github.io. 6/6


Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

Tables are a goldmine for accurate, fresh, domain data that LLMs should be grounded in for RAG, factver/QA, text2sql: retrieval is key! We introduce 🎯TARGET, a benchmark for evaluating table retrieval! Code,data,paper: target-benchmark.github.io Thinking bm25 ftw? Think twice..🧵

Tweet Image 1

Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

Check out our newest workshop paper: TARGET: Benchmarking Table Retrieval for Generative Tasks This marks my first academic publication, and I'm super grateful for the guidance from @MadelonHulsebos and @adityagp Thank you for making this such a rewarding and fun experience!

Tables are a goldmine for accurate, fresh, domain data that LLMs should be grounded in for RAG, factver/QA, text2sql: retrieval is key! We introduce 🎯TARGET, a benchmark for evaluating table retrieval! Code,data,paper: target-benchmark.github.io Thinking bm25 ftw? Think twice..🧵

Tweet Image 1


Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

🎦 Watch the recording of Madelon Hulsebos's seminar at BIDS! "There is actually a lot of structured data available on the web.... But really, the task here is that we should be able to retrieve that easily." #datascience #structureddata #dataretrieval bids.berkeley.edu/news/madelon-h…

Tweet Image 1

Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

Open-rank search (unusual for EECS) - please RT and apply!

EECS is hiring! Open faculty positions are now available. We welcome applicants from all areas focusing on originality and research promise. Join us in shaping the future of EECS! #UCBerkeley #EECS 🔗 More info: bit.ly/3AaLZdY 🔗 bit.ly/3YgWDb6



Madelon Hulsebos 🦋 madelonhulsebos.bsky.social Reposted

Our paper, "Data Void Exploits: Tracking & Mitigation Strategies," has just received the Best Paper Award at @cikm2024 ! 🏆 Data voids are gaps in online information, which are often exploit to spread disinformation. More details 👇 #CIKM2024 #DataVoids #Disinformation #KGs

Tweet Image 1

Loading...

Something went wrong.


Something went wrong.