The Archive

In high-stakes applications, reliable confidence estimates are as important as the predictions themselves. Confidence calibration ensures that predicted probabilities reflect the likelihood of correctness, making it essential for safe deployment of deep learning models. However, existing methods typically assume access to clean validation data, which is often unrealistic due to label noise and domain shifts. This thesis develops methods for improving calibration under these conditions. First, we address calibration under label noise. Standard methods can produce misleading confidence estimate...

AI notetaking hardware has taken off over the past couple of years, with credit-card-sized devices, pendants, pins, and even transcribing earbuds all promising to capture your meetings and turn them into summaries and action items. Now, a whole wave of wearables — rings especially — are betting people want to capture stray thoughts and ideas the same way. One of […]

We present RT-SEMamba, a fully causal speech enhancement (SE) model built upon causal time-frequency Mamba blocks. Unlike Transformer-based architectures that rely on a growing key-value cache, Mamba propagates a fixed-size recurrent state per layer, enabling memory- and bandwidth-efficient long-form inference. We further introduce a progressive knowledge distillation (KD) strategy that compresses an 8-layer teacher into a shallow 1-layer student by jointly distilling complex spectral outputs and intermediate representations. On Voicebank-DEMAND, the 8-layer RT-SEMamba achieves 3.32 PESQ with...

From the Pixel 11 series and a brand new competitor to Apple’s AirTag, here are all the announcements from the Made by Google 2026 event.

Rubric-based evaluators commonly treat rubrics as prompt context or flat criteria: they specify what to judge but leave criterion composition implicit, even when natural-language rules state it. We introduce Graph-Structured Rubrics (GSR), which compiles a rubric into a response-independent typed evaluation graph before observing responses. Criterion nodes elicit judgments; transformation, reduction, and gating operators compose them through named ports; and a task-specific output mapping, termed Readout, converts the unique sink into a score or preference. Compilation rejects malformed or ty...

Protein language models (PLMs) have transferred the latest advances from natural language processing to computational biology. These models, trained on large corpora of protein sequence data, are widely used to translate amino acid sequences into latent-space embeddings, ready for use in diverse downstream tasks (DTs). By a common consensus, embeddings from the model's last layer are used, and the model's internal behavior remains poorly understood. We analyzed 13 PLMs across 15 DTs from 11 datasets to investigate the informativeness of embeddings created in intermediate PLM layers. We traine...

Vision-language models, such as contrastive language-image pre-training (CLIP)-based approaches, have reached state-of-the-art (SOTA) results in medical artificial intelligence. However, recent work reveals that CLIP-based models remain vulnerable to shortcuts. We investigate how real-world shortcuts manifest across different layers of the medical CLIP-based model, MedCLIP, and its vision encoder, a frozen ResNet-50. We attach 17 linear classification probes to the intermediate layers of the ResNet-50 and train them on three different dataset configurations and targets: NIH-CXR14 (pneumothora...

We consider the setting of Normalizing flows with approximate inverses, an established paradigm spanning both full-dimensional ($d=D$) and bottleneck ($d<D$) settings, and group these models under the term flow autoencoders. We present a theoretical investigation into their training dynamics and prove that the proposed loss used by existing approaches is suboptimal; specifically, both encoder and decoder surrogates must be optimized in alignment with reconstruction loss. Guided by these insights, we propose Normalizing Autoencoder (NAE), which employs a novel conditional loss that aligns the ...

Introducing sign-language-to-text (SL2T), our breakthrough model powering new sign language features for Deaf and hard of hearing users.

At least there’s no new proprietary charger this year. Huzzah!! | Photo: David Imel / The Verge The $399 Google Pixel Watch 5 isn't about the hardware. Sure, there's a new satin pyrite case finish, a few new strap colors, and a Steph Curry Special Edition. Under the hood, there's a slightly faster Qualcomm processor and an itty-bitty battery bump. There's a $50 price hike from last year, too, because the Pixel Watch 5 isn't immune to RAMageddon - none of us are. Otherwise, no one would blame you for looking at this watch and thinking absolutely nothing's changed. That's because the big update...

Remember that much-hyped story about an Australian tech entrepreneur using ChatGPT, Grok, and other AI tools to craft a personalized cancer vaccine for his dog? Well, surprise: he's launched a startup. That entrepreneur is Paul Conyngham, who says he is launching Gamgee to offer "personalised mRNA cancer vaccines for dogs." But his ambitions go well beyond dogs - and pets - with Gamgee's website promising the company will harness AI and genetics to develop personalized treatments for a range of diseases and species, including humans. At the center of Gamgee's origin story is Rosie, Conyngham'...

You’ll have to be fine with letting Grok sign into your online accounts, however. | Image: SpaceXAI SpaceXAI has introduced Grok Bot, an always-on AI agent service designed to behave like independent "AI teammates" that can do your work for you. The bots share their own cloud-based computer environment, and can sign into apps, tools, and websites you already use to complete multi-step workplace tasks, only coming back when their assigned work is completed or if something requires approval. Grok Bot (or Bots, as SpaceXAI inconsistently pluralizes it) is the latest push from Elon Musk's AI comp...

Blacksmith says revenue has grown more than tenfold over the past year.

Anthropic is adding watermarking in response to the E.U.'s AI law. It's a terrible idea, first and foremost for philosophical reasons.

Speculative Decoding by any other name would distil as sweet

OpenAI research reveals how enterprises are adopting agentic AI, using ChatGPT and Codex, and how frontier firms are pulling ahead in AI adoption.

After a former lead writer claimed Saber "replaced me with ChatGPT," CEO Matthew Karch now claims, "Neither Saber nor Unigine have replaced any writers with AI," for the Rideshare "Stimulator" game announced last month, developed by Unigine. The writer, Stella Sacco, says differently, however, posting on Bluesky that "I was lead writer on this one! And Saber replaced me with ChatGPT midway through development. All the passenger voices were AI too. Either they changed direction at some point or they're not disclosing it on Steam." Saber describes Rideshare "Stimulator" as an "immersive driving...

See how RingCentral uses ChatGPT Work and Codex to accelerate AI product development and centralize operational intelligence across engineering and operations.

There are no lossless transformations of natural-language text Sophie Alpert shares her "internal policy on acceptable use of AI writing by engineers". It's a short read (supporting its own recommendations) and really good. If you chose to have LLMs help massage your writing the following rule seems crucial to me: You must stand behind every idea and every sentence in your docs . It is your responsibility to make sure that the entire document is representative of your own thoughts before you share it. If a reviewer asks, “What did you mean by this line?”, it’s not acceptable to reply with “Oh...

Stealing Reasoning Traces from Proprietary LLM APIs A vanity domain name ( stolen-thoughts.com ) for a neat paper : Anthropic, OpenAI, and Google return encrypted chain-of-thought blocks to clients that can be replayed across sessions, users, and models. We take a trace produced by a frontier model, replay it into a weaker sibling, jailbreak the weaker model, and recover the stronger model’s hidden reasoning in plaintext You can see an example of these encrypted blocks by running: curl https://api.openai.com/v1/responses \ -H " Content-Type: application/json " \ -H " Authorization: Bearer $( ...

The U.S. VC firm still has more than 55% of its previous $650 million India fund available for deployment.

Pharma is suddenly paying for Bio × AI tools, and Chai is leading the pack with four deals closed this summer. Cofounder Matt McPartlon and Product leader Neil Patil explain why.

Release: datasette-upload-dbs 0.5a0 This plugin has been around for a while - it lets users upload a brand new SQLite database to a hosted Datasette instance, at which point that database will start being served by that instance. It can also be used to atomically swap a database with a more recent version. The uploaded database is saved to a file, verified, then swapped in so /name starts serving the new one. The new release adds a formalized API, so you can replace an existing database (or add a new one) like this: curl -X POST \ -H "Authorization: Bearer $API_TOKEN" \ -H "Accept: applicatio...

But will Gemini's surge survive slowing model releases?

OpenAI is finally bringing a dedicated ChatGPT desktop app to Linux operating systems.

Video is a core data path across NVIDIA Jetson applications, from robotics and intelligent video analytics to industrial automation, healthcare, media... Video is a core data path across NVIDIA Jetson applications, from robotics and intelligent video analytics to industrial automation, healthcare, media processing, and remote operations. A system may capture several cameras, decode network streams, run AI inference or conventional vision processing, draw results, and encode video for storage or delivery. The individual calls are… Source

Gemini is keeping pace with OpenAI’s ChatGPT, which hit 1 billion monthly active users back in June.

Learning reliable surgical manipulation policies is bottlenecked by the scarcity of action-labeled demonstrations: teleoperated surgical robot (e.g., dVRK) trajectories with synchronized kinematics are costly to collect, while surgical tasks demand precise contact handling, long-horizon reasoning, and bimanual coordination. Endoscopic video is comparatively inexpensive and abundant relative to synchronized video--kinematics trajectories, and a natural way to exploit it is to learn world models of surgical scenes. However, existing surgical world models use video primarily for simulation or po...

Synthetic dialogue generation offers a way to study conversational dynamics in sensitive domains where real data are difficult to access, release, or annotate. The underlying abuse may occur online or offline: threats and coercion can appear directly in messages, while behaviours such as surveillance, isolation, stalking, and physical violence may be planned, disclosed, or referred to conversationally. Privacy and legal constraints make it difficult the release of large-scale real conversation datasets; existing work has mostly focused on sentence-level toxicity of online abuses, leaving a ga...