Biagio La Rosa

PostDoc working on Explainable Deep Learning @ucsc | PhD @SapienzaRoma

[Biagio La Rosa @larosabiagio]

I am happy to share that our paper "Guaranteed Optimal Compositional Explanations" (with @leilanigilpin) has been accepted at ICML 2026 as an Oral Paper.

We will present it in Seoul on July 7th. If you're attending, I'd be happy to exchange ideas there! 🧵(1/8)

[Biagio La Rosa @larosabiagio]

Fortunately, after many iterations (and many nights), I found a solution that worked. I would like to thank @leilanigilpin for giving me the freedom to think, explore, and take risks. As a prospective PI, I hope to offer the same environment to my future students.

8/8

[Biagio La Rosa @larosabiagio]

I am happy to share that our paper "Guaranteed Optimal Compositional Explanations" (with @leilanigilpin) has been accepted at ICML 2026 as an Oral Paper.

We will present it in Seoul on July 7th. If you're attending, I'd be happy to exchange ideas there! 🧵(1/8)

[Biagio La Rosa @larosabiagio]

A more personal note. I had this idea in the back of my mind for years. It was one of those high-risk, high-reward projects that could easily have turned into months of equations leading nowhere.
7/8

[Peter Richtarik @peter_richtarik]

I am an AC for ICLR 2026. One of the papers in my batch was just withdrawn. The authors wrote a brief response, explaining why the reviewers failed at their job. I agree with most of their comments. The authors gave up. They are fed up. Just like many of us. I understand. We pretend the emperor has clothes, but he is naked.

Here is the final part of their withdrawal notice. I took the liberty to make it public, to highlight that what we are doing with AI conference reviews these last few years is, basically, madness.

---

Comment: We thank the reviewers for their time.

However, upon reading the reviews for our paper, it became immediately apparent that the four "reject" ratings are not based on good-faith academic disagreement, but on a critical failure to read the submitted paper.

The reviews are rife with demonstrably false claims that are directly contradicted by the text. The core justifications for rejection rely on asserting that key components are "missing" when they are explicitly detailed in the manuscript. Some specific examples are (and many are even fake claims).

Claim: Harder tasks like GSM8K are missing.

Fact: GSM8K results are in many tables, like Table 2 (Section 4.2) and Appendix G.

Claim: The method does not use per-layer ranks.

Fact: This is the entire point of our method. The reviewer clearly mistook our method for the baselines. (Section 2, Table 1).

Claim: The GP kernel is not specified.

Fact: It is specified in Appendix E (Table 6).

Claim: There is no ablation of the method's three stages.

Fact: Section 4.4 ("Ablation Study") and Appendix J are dedicated to this.

Reviewers have a fundamental responsibility to read and evaluate the work they are assigned. The nature of these errors is so fundamental, so systemic in overlooking explicit content, that it goes far beyond what "limited time" or "oversight" can explain. This work has gone through several rounds of revision over the last year. In earlier submissions, the paper usually received borderline or weak-accept scores.

Numerous signs strongly suggest that some reviewers are relying entirely on AI tools to automatically generate peer reviews, rather than fulfilling their fundamental responsibility of personally reading and evaluating manuscripts.

We strongly protest this.

This is a gross disrespect to the authors. It is a flagrant desecration of the reviewer's sacred duty. It fundamentally undermines the integrity of the entire peer-review process.

Given that the reviews are not based on the actual content of our paper, we have decided to withdraw the submission.

We leave this comment so that future readers of the OpenReview page are aware that the items described as "missing" are already present in the submitted manuscript. These negative reviews for this submission are factually unsound and do not reflect the content of the paper. We cannot and will not accept an assessment that is not based on the work we actually submitted.

[Biagio La Rosa @larosabiagio]

If you are located in the area or interested in the topic, let's chat and connect!! Can't wait to see what the future holds! 2/2

[Biagio La Rosa @larosabiagio]

I’m happy to announce the exciting next step of my academic career: I’m moving to the US to join @ucsc as a postdoctoral researcher! I’ll be working at the AIEA lab, led by the amazing Prof. @leilanigilpin and continue my research on explaining deep neural networks and AI! 1/2>>

[Leilani sigmoid.social/@leilani @leilanigilpin]

A nice write up on our XAI for DRL workshop at AAAI earlier this year with @larosabiagio @runnerdude97
@webrot @spideralessio @MProietti_

[Biagio La Rosa @larosabiagio]

Heading to #AAAI2024 in a few hours!! If anyone is around drop me a message on Whova and we can meet there! ps: visit our workshop in the quoted message!

[Leilani sigmoid.social/@leilani @leilanigilpin]

Heading to #AAAI2024 today with @runnerdude97 to meet up with @larosabiagio and others for the XAI for DRL workshop: https://t.co/8BWfzgwIec. Let me know if you're around!

[Biagio La Rosa @larosabiagio]

@leilanigilpin Explaining the explanations?! There is something that we missed!

[Biagio La Rosa @larosabiagio]

I and @leilanigilpin are going to present the poster of this work tomorrow morning (10:45am) at #NeurIPS2023 . Paper #1522. Come and chat together about it!

[Biagio La Rosa @larosabiagio]

Interested in understanding what neurons of #DeepNeuralNetworks are able to recognize at different ranges of activations? Thrilled to announce that our #XAI paper has been accepted at #NeurIPS2023 (https://t.co/XRD596uCcz) and it is going to be presented as a poster in 🧵1/5 >>

[Biagio La Rosa @larosabiagio]

@berkustun Hi, it seems that the current form to send the application doesn't include any field for the cover letter. It is required for the application, isn't it?

[Biagio La Rosa @larosabiagio]

Please retweet! We extended the deadline for our #XAI workshop XAI4DRL@AAAI2024 to Nov 21st! https://t.co/MpOtIt3qj8 Any XAI paper is welcome, even if RL is not involved! Please check the CFP and FAQ to know more about it!

[Biagio La Rosa @larosabiagio]

We have in mind several directions to further investigate the topic in the future and we are excited about it, so we encourage and welcome any feedback or exchange of ideas on the paper’s topic! #deeplearning #explainableAI #interpretableML /5

[Biagio La Rosa @larosabiagio]

Interested in understanding what neurons of #DeepNeuralNetworks are able to recognize at different ranges of activations? Thrilled to announce that our #XAI paper has been accepted at #NeurIPS2023 (https://t.co/XRD596uCcz) and it is going to be presented as a poster in 🧵1/5 >>

[Biagio La Rosa @larosabiagio]

progressive specialization, and activation polysemy. This paper is a joint work with @webrot and @leilanigilpin . We released the code at https://t.co/Al4HpwUwUf. 🧵4/5