@ExplainableNL profile picture
explAInable.NL @ExplainableNL
Joined October 2019
  • Tweets 251
  • Following 219
  • Followers 423
5 Photos and videos
Exciting results from StanfordNLP (with D'Oosterlinck from Gent) on Causal Proxies: using symbolic surrogate models for interpreting deep learning, and testing for causality using counterfactual interventions.
Karel
@KarelDoostrlnck
30 Sep 2022
🚨Preprint🚨 Interpretable explanations of NLP models are a prerequisite for numerous goals (e.g. safety, trust). We introduce Causal Proxy Models, which provide rich concept-level explanations and can even entirely replace the models they explain. arxiv.org/abs/2209.14279 1/7

ALT A factual restaurant review is supplemented with a counterfactual review, where one concept is changed. A third review is sampled according to this concept change. Using the factual input and third input, a Causal Proxy Model is trained to predict the output of a black-box model on the counterfactual input. This training is done using an intervention on hidden representations of the Causal Proxy Model, forcing the model to localize concept-level information in hidden states.
explAInable.NL retweeted
Marco Virgolin πŸ‡ΊπŸ‡¦@MarcoVirgolin
6 Jul 2022
Symbolic regression (SR) is the problem of finding an accurate model of the data in the form of a (hopefully elegant) mathematical expression. SR has been thought to be hard and traditionally attempted using evolutionary algorithms. This begs the question: is SR NP-hard? 1/2
3
43
265
Excited to see our work featured in the Fortune. Thank you so much, @jeremyakahn!
Jeremy Kahn@jeremyakahn
22 Mar 2022
What’s wrong with β€œexplainable" artificial intelligence? Pretty much everything. fortune.com/2022/03/22/ai-ex…
1
1
11
explAInable.NL retweeted
Stanford HAI
@StanfordHAI
24 Mar 2022
This year's Spring Conference focuses on foundation models, accountable AI, and embodied AI. HAI Associate Director and event co-host @chrmanning explains these key areas and why you should not miss this event: stanford.io/3IxnjdH
8
17
explAInable.NL retweeted
Fernando P. Santos@fernandopsantos
24 Mar 2022
Interested in Explainable AI and Finance? Check out this opportunity for a Tenure Track Assistant Professor position at the Informatics Institute, University of Amsterdam! Deadline extended to 3 April 2022.
This tweet is unavailable
2
6
explAInable.NL retweeted
Marco Virgolin πŸ‡ΊπŸ‡¦@MarcoVirgolin
24 Mar 2022
And happy that also our work "On genetic programming representations and fitness functions for interpretable dimensionality reduction" made it to @GeccoConf! Preprint: arxiv.org/abs/2203.00528 A short explanation πŸ‘‡ 1/8

On genetic programming representations and fitness functions for...

Dimensionality reduction (DR) is an important technique for data exploration and knowledge discovery. However, most of the main DR methods are either linear (e.g., PCA), do not provide an explicit...

arxiv.org
Marco Virgolin πŸ‡ΊπŸ‡¦@MarcoVirgolin
24 Mar 2022
Happy to share that our work "Evolvability Degeneration in Multi-Objective Genetic Programming for Symbolic Regression" has been accepted at @GeccoConf! πŸ₯³πŸͺ…πŸΎ Preprint: arxiv.org/abs/2202.06983. A high-level🧡 of what's going on hereπŸ‘‡ 1/8
2
2
5
explAInable.NL retweeted
Nils Trost@TrostNils
24 Mar 2022
I visualized my last #semantle game with a UMAP of the word embeddings. Here's the result: bp.bleb.li/viewer?p=D5d3y Semantle is a word guessing game by @NovalisDMT where your guesses, unlike in #wordle, are ranked by their similarity in meaning, not spelling, to the secret word.
6
9
51
explAInable.NL retweeted
NL4XAI@NL4XAI
21 Mar 2022
πŸ“’#MSCAJobAlert Last days to apply to the PhD student position in #AI within @NL4XAI @MSCActions at @citiususc, ES. Join us and work on the following topic: From Grey-box Models to Explainable Models. βŒ›οΈDeadline 31/03/2022 ApplyπŸ‘‰nl4xai.eu/open_position/esr2… @EU_H2020
9
11
explAInable.NL retweeted
DIGIT πŸ‡ͺπŸ‡Ί
@EU_DIGIT
24 Mar 2022
πŸ“’Call for contributions to help identify Europe’s most Critical #OpenSourceSoftware ! We urge all national, regional and local public administrations across all EU 27 member states, to participate! Learn moreπŸ‘‰europa.eu/!HXxQqp #FOSSEPS #ThinkOpen
24
21
Interesting recent trend in XAI: serious efforts to study the interaction between explainability methods and the *people* who the explanations are addressed to. Jacovi et al. develop a framework to link interpretability tools to things already known about human cognition. 1/3
Alon Jacovi@alon_jacovi
31 Jan 2022
What does it mean for a user to "understand" the explanation of an AI's decision? New paper πŸ™Œ arxiv.org/abs/2201.11239 Diagnosing AI Explanation Methods with Folk Concepts of Behavior w\ @BastingsJasmijn @sebgehr @yoavgo @fajtak Digest:🧡
1
4
16
more replies
Krishna et al. study what practioners *do* in practice with what posthoc interpretability tools produce -- and with the strong disagreements between those tools. *tldr*: XAI is here to stay, but its methods need to be handled with care! 3/3 x.com/hima_lakkaraju/status/…

π™·πš’πš–πšŠ π™»πšŠπš”πš”πšŠπš›πšŠπš“πšž
@hima_lakkaraju
7 Feb 2022
Excited to share our work "The Disagreement Problem in Explainable Machine Learning" arxiv.org/pdf/2202.01602.pdf. We present new results showing that post hoc explanation methods often disagree in practice & practitioners resolve those disagreements using arbitrary heuristics [1/N]
1
2
And from a few months ago, Ehsan (2021) and team, including @mark_riedl, highlight how the AI background of recipients of explanations influences their interpretations. 4/3 mobile.twitter.com/UpolEhsan…

Upol Ehsan@UpolEhsan
30 Jul 2021
🚨 New pre-print alert! 🚨 Excited to share β€œThe Who in Explainable AI: How AI Background Shapes Perceptions of AI Explanations” w/ the amazing team: @samirpassi,@QVeraLiao,Larry Chan,Ethan Lee,@michael_muller,@mark_riedl πŸ”—arxiv.org/abs/2107.13509 πŸ’‘Findings at a glance... 1/n
1
"Transparency and explainability pertain to the technical domain ... leaving the ethics and epistemology of AI largely disconnected. In this talk, Russo will focus on how to remedy this problem and introduce an epistemology for glass box AI that can explicitly incorporate values"
Rineke Verbrugge@RinekeV
8 Feb 2022
Lecture by Federica Russo @federicarusso: Connecting the ethics and epistemology of AI. This Thursday 10 Feb, 12-13 h CET, online. Moderated by AybΓΌke Γ–zgΓΌn. For more information and the way to get access, see: uva.nl/en/shared-content/sub…
A team of researchers from Amsterdam and Rome proposes CF-GNNExplainer: an explainability method for the popular Graph Neural Networks. The method iteratively removes edges from the graph, returning the minimal perturbation that leads to a change in prediction.
Ana LučiΔ‡@__alucic
18 Jan 2022
Excited that our paper with @maartjeterhoeve, @gtolomei, @mdr and @fabreetseo on counterfactual explanations for GNNs has been accepted to #AISTATS2022!!! Preprint available here: bit.ly/3If7Hfd πŸ₯³
2
6
#explainableAI #fairness
This tweet is unavailable
1
Interesting blogpost on SHAP / feature attribution using Shapley values, by researchers from Dutch medical AI company Pacmed.
Pacmed@Pacmedhealth
17 Jan 2022
At Pacmed we care about improving medical practice with the help of AI. We often use tree-based models 🌳 in combination with SHAP values to gain a better understanding of what models do. But... which version of SHAP is best to use? 1/3 pacmedhealth.medium.com/expl…
explAInable.NL retweeted
Badr M. Abdullah, PhD πŸ‡ΎπŸ‡ͺ@badr_nlp
17 Nov 2021
Replying to @gchrupala
A paper in ICASSP 2020 proposed probing by "audification" of hidden representations in ASR model. They learn a speech synthesizer on top of the ASR representations. They have a nice video of their work here youtu.be/6gtn7H-pWr8

1
1
3
This paper analyzing discrete representations in models of spoken language with @bertrand_higy @liekegelderloos and @afraalishahi will appear at #BlackboxNLP #EMNLP2021 arxiv.org/abs/2105.05582

Discrete representations in neural models of spoken language

The distributed and continuous representations used by neural networks are at odds with representations employed in linguistics, which are typically symbolic. Vector quantization has been proposed...

arxiv.org
1
3
16
explAInable.NL retweeted
Yonatan Belinkov@boknilev
11 Nov 2021
Replying to @sarahookr @anmarasovic
Hot take from @wzuidema : progress in probing classifiers will not come from sophisticated probing techniques but from the hard work of forming better hypotheses.
3
12
Load more