Daniel Arteaga @dnlrtg.bsky.social · 26/11/2025The question isn't whether AI doom is likely. It's whether the expected harm is significant enough to act on. Given the math? The answer is clearly yes. We don't need certainty to justify precaution. We need responsible risk assessment. (7/7) 020
Daniel Arteaga @dnlrtg.bsky.social · 26/11/2025This same risk framework (expected value of harm) applies to non-catastrophic AI risks too: - Algorithmic bias - Economic displacement - Privacy violations - Misinformation at scale If the product of probability and harm that matters. (6/7) www.ibm.com/think/insigh...ibm.com10 AI dangers and risks and how to manage them | IBMA closer look at 10 dangers of artificial intelligence and actionable risk management strategies to consider today. 130
Daniel Arteaga @dnlrtg.bsky.social · 26/11/2025Most AI researchers agree existential risk from AI has LOW probability. But low ≠ zero. And when we're talking about existential outcomes, non-zero probabilities require: - Serious research - Robust safety measures - Contingency planning (5/7) en.wikipedia.org/wiki/Safety_...en.wikipedia.orgSafety engineering - Wikipedia 120
Daniel Arteaga @dnlrtg.bsky.social · 26/11/2025We have precedent for this thinking. Before launching the Large Hadron Collider, CERN seriously studied scenarios like micro black holes destroying Earth. Did physicists think it would happen? No. But the potential harm was so large, they HAD to investigate. (4/7) en.wikipedia.org/wiki/Strange...en.wikipedia.orgStrangelet - Wikipedia 100
Daniel Arteaga @dnlrtg.bsky.social · 26/11/2025If potential harm approaches infinity, only a truly negligible probability makes the overall risk acceptable. Low probability ≠ no risk when the stakes are existential. (3/7) en.wikipedia.org/wiki/Risk_as...en.wikipedia.orgRisk assessment - Wikipedia 120
Daniel Arteaga @dnlrtg.bsky.social · 26/11/2025What matters is: Probability × Magnitude of Harm = Risk When harm could be civilization-ending, even tiny probabilities demand serious attention. (2/7) en.wikipedia.org/wiki/Risk_ma...en.wikipedia.orgRisk matrix - Wikipedia 100
Daniel Arteaga @dnlrtg.bsky.social · 26/11/2025The AI existential risk debate is often framed wrong. People focus on "how likely is this?" But that's not the right question. (1/7) 🧵 en.wikipedia.org/wiki/Existen...en.wikipedia.orgExistential risk from artificial intelligence - Wikipedia 110
Daniel Arteaga @dnlrtg.bsky.social · 14/11/2025🏆 It's a pleasure to announce the winner of the Dolby Barcelona Scientific Paper Award 2025: Guillem Cortès Sebastià for PeakNetFP, a scalable neural audio fingerprinting system robust to extreme time stretching. Details: professional.dolby.com/legal/barcel...professional.dolby.comDolby Barcelona Scientific Paper Award 2025 - Dolby Professional 001
Daniel Arteaga @dnlrtg.bsky.social · 27/10/2025Yet in AI research (which is essentially statistical modeling)we routinely abandon these basic practices. The irony is striking. 100
Daniel Arteaga @dnlrtg.bsky.social · 27/10/2025In other scientific fields (natural and social sciences), proper statistical analysis is fundamental. You simply cannot publish without it. 100
Daniel Arteaga @dnlrtg.bsky.social · 27/10/2025There's also the added problem that metrics often don't correlate with perception. A 0.1 dB SDR improvement might be meaningless perceptually. But this issue has been discussed more often than the statistical rigor problem. 110
Daniel Arteaga @dnlrtg.bsky.social · 27/10/2025❌ Claims of "superior performance" based on point estimates alone Example: Paper A reports 15.21 dB, Paper B reports 15.01 dB. Is this difference meaningful or just noise? Do those decimal places have any meaning? Usually impossible to tell from the paper. 100
Daniel Arteaga @dnlrtg.bsky.social · 27/10/2025❌ Values without error bars/confidence intervals ❌ Standard deviations sometimes quoted but no uncertainty estimates of means ❌ No significance testing whatsoever ❌ No effect size analysis ❌ No exploratory analysis beyond the mean 100
Daniel Arteaga @dnlrtg.bsky.social · 27/10/2025We're not even applying methods from first-year undergraduate physics—like reporting results with error bars. The problems I regularly see would make any physics professor cringe. 100
Daniel Arteaga @dnlrtg.bsky.social · 27/10/2025Coming from a physics background and being familiar with research methodology in fields like neuroscience, I'm struck by the lack of statistical practices in ML/AI papers. Most papers fail at basic statistical analysis that would be expected in other scientific fields. 🧵 110
Daniel Arteaga @dnlrtg.bsky.social · 13/10/2025At @waspaa.com in Lake Tahoe 🏔️ — great talks, great atmosphere, and maybe a bear or two 🐻. Happy to present today our work on generative AI for RIR! 000
Daniel Arteaga @dnlrtg.bsky.social · 12/09/2025Dolby Barcelona Scientific Paper Award 2025 - deadline extended to 30th September due to ICASSP paper submission overlap. Final call! Details: professional.dolby.com/legal/barcelona-scientific-paper-award-2025/professional.dolby.comDolby Barcelona Scientific Paper Award 2025 - Dolby Professional 000
Daniel Arteaga @dnlrtg.bsky.social · 04/09/2025Just 11 days left until the Sept 15 deadline and we've already received some excellent submissions. Still working on yours? Don't wait: showcase your 2024-2025 sound research now. 000
Reposted by Daniel ArteagaDaniel Arteaga @dnlrtg.bsky.social · 18/07/2025New paper! We're introducing a new way to generate realistic room impulse responses not from room geometry, but by directly controlling acoustic parameters like reverb time and direct-to-reverb ratio. 🔗 Demo: silviaarellanogarcia.github.io/rir-acoustic/ 📄 Paper: arxiv.org/pdf/2507.12136 122
Daniel Arteaga @dnlrtg.bsky.social · 18/07/2025This work was the result of Silvia Arellano's internship in Dolby Barcelona with us. Come explore the demo here: 🔗 silviaarellanogarcia.github.io/rir-acoustic/ 📄 Paper: arxiv.org/pdf/2507.12136 Feedback & questions welcome! 000
Daniel Arteaga @dnlrtg.bsky.social · 18/07/2025We explore 4 DAC-based models: 1️⃣ AR w/ cross-attention 2️⃣ AR w/ classifier guidance 3️⃣ MaskGIT w/ adaptive layer norm 4️⃣ Flow matching The MaskGIT model achieves the best subjective quality (avg. 70 MUSHRA score), beating state of the art comparisons. 100
Daniel Arteaga @dnlrtg.bsky.social · 18/07/2025Instead of simulating room geometry, we train four different generative model to produce RIRs conditioned on acoustic attributes (T30, T15, EDT, D50, C80, source-receiver distance) 100
Daniel Arteaga @dnlrtg.bsky.social · 18/07/2025New paper! We're introducing a new way to generate realistic room impulse responses not from room geometry, but by directly controlling acoustic parameters like reverb time and direct-to-reverb ratio. 🔗 Demo: silviaarellanogarcia.github.io/rir-acoustic/ 📄 Paper: arxiv.org/pdf/2507.12136 122
Reposted by Daniel ArteagaDaniel Arteaga @dnlrtg.bsky.social · 18/06/2025I'm happy to share that in the Dolby office in Barcelona we are offering an award for outstanding scientific papers in sound research. Open to students and early-career researchers with connections to Catalonia. Deadline 15th September. professional.dolby.com/legal/barcel...professional.dolby.comDolby Barcelona Scientific Paper Award 2025 - Dolby Professional 101
Daniel Arteaga @dnlrtg.bsky.social · 18/06/2025Last winner, Eloi Moliner, pioneered diffusion models in AI for audio restoration. Could you be next? arxiv.org/abs/2210.15228arxiv.orgSolving Audio Inverse Problems with a Diffusion ModelThis paper presents CQT-Diff, a data-driven generative audio model that can, once trained, be used for solving various different audio inverse problems in a problem-agnostic setting. CQT-Diff is a neu... 000
Daniel Arteaga @dnlrtg.bsky.social · 18/06/2025I'm happy to share that in the Dolby office in Barcelona we are offering an award for outstanding scientific papers in sound research. Open to students and early-career researchers with connections to Catalonia. Deadline 15th September. professional.dolby.com/legal/barcel...professional.dolby.comDolby Barcelona Scientific Paper Award 2025 - Dolby Professional 101
Daniel Arteaga @dnlrtg.bsky.social · 28/05/2025The key issue isn't the most likely outcome — it's the worst-case scenario we must be prepared for. arxiv.org/abs/2401.02843 000
Daniel Arteaga @dnlrtg.bsky.social · 28/05/2025Just as nuclear research is subject to international oversight, frontier AI development should be too. We need strong global regulatory frameworks for models with potentially vast power. www.iaea.orgiaea.orgInternational Atomic Energy Agency | Atoms for Peace and DevelopmentThe IAEA is the world's centre for cooperation in the nuclear field, promoting the safe, secure and peaceful use of nuclear technology. It works in a wide range of areas including energy generation, h... 100
Daniel Arteaga @dnlrtg.bsky.social · 28/05/2025Anthropic deserves credit for its serious commitment to LLM safety — but where are similar efforts from other big tech players like Google or OpenAI? 🔗 www.anthropic.com/news/activat...anthropic.comActivating AI Safety Level 3 ProtectionsWe have activated the AI Safety Level 3 (ASL-3) Deployment and Security Standards described in Anthropic’s Responsible Scaling Policy (RSP) in conjunction with launching Claude Opus 4. The ASL-3 Secur... 100
Daniel Arteaga @dnlrtg.bsky.social · 27/05/2025Example audio samples here: metlosz.github.io/dealiasing_a...metlosz.github.ioDeep learning based spatial aliasing reduction in beamforming for audio capture 000
Daniel Arteaga @dnlrtg.bsky.social · 27/05/2025Spatial aliasing occurs when microphone spacing in an array is too large relative to the wavelength of sound, degrading the accuracy of beamforming. As far as we know, this is the first deep learning paper that addresses directly this problem (other approaches deal with this in an indirect way). 100
Daniel Arteaga @dnlrtg.bsky.social · 27/05/2025Just dropped our new paper on arXiv (accepted for Interspeech 2025) Mateusz Guzik (while interning at Dolby) developed together with us a cool deep learning approach that fixes spatial aliasing in beamforming. Paper: arxiv.org/abs/2505.19781arxiv.orgDeep learning based spatial aliasing reduction in beamforming for audio captureSpatial aliasing affects spaced microphone arrays, causing directional ambiguity above certain frequencies, degrading spatial and spectral accuracy of beamformers. Given the limitations of conventiona... 110
Daniel Arteaga @dnlrtg.bsky.social · 19/05/2025Updated my lecture notes in Ambisonics — now available on Zenodo. They started as course material but have helped quite a few people beyond that. Not a textbook, just some pretty basic lecutre notes. Feedback welcome! zenodo.org/records/1546...zenodo.orgIntroduction to AmbisonicsThese are a preliminary version of some lecture notes on Ambisonics, prepared for the course “Àudio 3D” at the Escola Superior Politècnica of Universitat Pompeu Fabra. Many elements are missing or can... 020
Reposted by Daniel ArteagaDaniel Arteaga @dnlrtg.bsky.social · 17/01/2025Are you a PhD student specializing in audio or AI? Interested in an internship with us in Barcelona? Apply here! We've had the pleasure of working with talented interns in the past, and the experience has been mutually very rewarding. www.linkedin.com/jobs/view/40...linkedin.comDolby Laboratories hiring AI PhD Intern, Sound Experiences Lab in Barcelona, Catalonia, Spain | LinkedInPosted 10:05:54 AM. Join the leader in entertainment innovation and help us design the future. At Dolby, science meets…See this and similar jobs on LinkedIn. 082
Daniel Arteaga @dnlrtg.bsky.social · 17/01/2025Are you a PhD student specializing in audio or AI? Interested in an internship with us in Barcelona? Apply here! We've had the pleasure of working with talented interns in the past, and the experience has been mutually very rewarding. www.linkedin.com/jobs/view/40...linkedin.comDolby Laboratories hiring AI PhD Intern, Sound Experiences Lab in Barcelona, Catalonia, Spain | LinkedInPosted 10:05:54 AM. Join the leader in entertainment innovation and help us design the future. At Dolby, science meets…See this and similar jobs on LinkedIn. 082
Reposted by Daniel ArteagaKeenan Crane @keenancrane.bsky.social · 09/12/2024Entropy is one of those formulas that many of us learn, swallow whole, and even use regularly without really understanding. (E.g., where does that “log” come from? Are there other possible formulas?) Yet there's an intuitive & almost inevitable way to arrive at this expression. 22543128
Daniel Arteaga @dnlrtg.bsky.social · 18/11/2024I also have a LinkedIn profile: www.linkedin.com/in/daniel-ar...linkedin.comDaniel Arteaga - Dolby Laboratories | LinkedInCurrently, I am a Senior Staff Researcher at Dolby Labs in Barcelona and an Associate… · Experience: Dolby Laboratories · Education: Universitat de Barcelona · Location: Greater Barcelona Metropolitan Area · 500+ connections on LinkedIn. View Daniel Arteaga’s profile on LinkedIn, a professional community of 1 billion members. 000
Daniel Arteaga @dnlrtg.bsky.social · 18/11/2024This is a continuation of my previous Twitter/X account, which I will keep alive but inactive for the time being: x.com/dnlrtgx.comx.com 100
Reposted by Daniel ArteagaJonathan Le Roux @jonathanleroux.bsky.social · 18/11/2024I initiated a starter pack for Audio ML. Let me know if you'd like to be added/removed. go.bsky.app/LGmct4z 466822
Daniel Arteaga @dnlrtg.bsky.social · 18/11/2024My first post in Bluesky! This is my new Bluesky feed where I intent to discuss things about physics, deep learning & AI, audio, research, technology and society, mostly in English. 240