Richard the Brave 🇨🇦Somehow have been a Mac user since the early 2000s, but had never used @flyingmeat’s Acorn image editor before yesterday. Bought it on a whim, no trial testing, and within *moments* of using knew it was the right choice to finally get on board. 🫡
Ctrl-OA-Reset 
MT Group at FBKOur pick of the week by @apierg: "Robust Pronoun Use Fidelity with English LLMs: Are they Reasoning, Repeating, or Just Biased?" by @dippedrusk, @lauscher, et al., 2024.
Great_AlbumsHere is my review of the intriguing new album by avant-pop artist Jane Weaver, published today in #PopMatters
https://www.popmatters.com/jane-weaver-love-constant-spectacle
Here is my review of Ride's Interplay, the latest album by the reunited British shoegaze band. Published today in PopMatters:
https://www.popmatters.com/ride-interplay-music-review
#RideBand #Shoegaze #Music #PickOfTheWeek
#PopMatters #CurrentSpins
Here is my review of Audio Vertigo, the new album by the British band Elbow. Published today in PopMatters.
Strong contender for my list of the year's best.
https://www.popmatters.com/elbow-audio-vertigo-review
#Elbow #GuyGarvey #Music #PickOfTheWeek #PopMatters #CurrentSpins
Here is my review of Bruce Dickinson's new solo album, The Mandrake Project, published today in PopMatters: https://www.popmatters.com/bruce-dickinson-mandrake-project-review
Our pick of the week by Lina Conti: "What is Interpretability?" by Erasmus, Tyler D. P. Brunet, and Eyal Fisher, 2020.
https://link.springer.com/article/10.1007/s13347-020-00435-2
What is Interpretability? - Philosophy & Technology
We argue that artificial networks are explainable and offer a novel theory of interpretability. Two sets of conceptual questions are prominent in theoretical engagements with artificial neural networks, especially in the context of medical artificial intelligence: (1) Are networks explainable, and if so, what does it mean to explain the output of a network? And (2) what does it mean for a network to be interpretable? We argue that accounts of “explanation” tailored specifically to neural networks have ineffectively reinvented the wheel. In response to (1), we show how four familiar accounts of explanation apply to neural networks as they would to any scientific phenomenon. We diagnose the confusion about explaining neural networks within the machine learning literature as an equivocation on “explainability,” “understandability” and “interpretability.” To remedy this, we distinguish between these notions, and answer (2) by offering a theory and typology of interpretation in machine learning. Interpretation is something one does to an explanation with the aim of producing another, more understandable, explanation. As with explanation, there are various concepts and methods involved in interpretation: Total or Partial, Global or Local, and Approximative or Isomorphic. Our account of “interpretability” is consistent with uses in the machine learning literature, in keeping with the philosophy of explanation and understanding, and pays special attention to medical artificial intelligence systems.
Our pick of the week by Beatrice Savoldi: "Measuring machine learning harms from stereotypes: requires understanding who is being harmed by which errors in what ways" by @ang3linawang, @baixuechunzi, @s010n, and @sulin_blodgett, 2024.
Our pick of the week by @apierg: "Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation" by Xu et al.
Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation
Moderate-sized large language models (LLMs) -- those with 7B or 13B parameters -- exhibit promising machine translation (MT) performance. However, even the top-performing 13B LLM-based translation models, like ALMA, does not match the performance of state-of-the-art conventional encoder-decoder translation models or larger-scale LLMs such as GPT-4. In this study, we bridge this performance gap. We first assess the shortcomings of supervised fine-tuning for LLMs in the MT task, emphasizing the quality issues present in the reference data, despite being human-generated. Then, in contrast to SFT which mimics reference translations, we introduce Contrastive Preference Optimization (CPO), a novel approach that trains models to avoid generating adequate but not perfect translations. Applying CPO to ALMA models with only 22K parallel sentences and 12M parameters yields significant improvements. The resulting model, called ALMA-R, can match or exceed the performance of the WMT competition winners and GPT-4 on WMT'21, WMT'22 and WMT'23 test datasets.