Nonlinear independent component analysis (ICA) provides an appealing framework for unsupervised feature learning, but the models proposed so far are not identifiable. Here, we first propose a new intuitive principle of unsupervised deep learning from time series which uses the nonstationary structure of the data. Our learning principle, time-contrastive learning (TCL), finds a representation which allows optimal discrimination of time segments (windows). Surprisingly, we show how TCL can be related to a nonlinear ICA model, when ICA is redefined to include temporal nonstationarities. In particular, we show that TCL combined with linear ICA estimates the nonlinear ICA model up to point-wise transformations of the sources, and this solution is unique --- thus providing the first identifiability result for nonlinear ICA which is rigorous, constructive, as well as very general.

Plus: Asus exits the smartphone market, and Sony partners with TCL on TVs. After delaying its Siri improvements to 2026, Apple's artificial intelligence plans are starting to take shape, at least according to the rumor mill. Bloomberg reports that Apple is turning Siri into a chatbot that will replace the voice assistant's existing interface, akin to OpenAI's ChatGPT. Codenamed Campos, the chatbot will be powered by Google's Gemini models and will be integrated into the iPhone, Mac, and iPad in their respective operating system updates later this fall. We'll likely learn more about Campos at Apple's developer event, WWDC, which usually takes place in June.

Hot on the tail of Amazon's Kindle Scribe Colorosoft, TCL is introducing its own take on a distraction-free note-taking and reading device. The TCL Note A1 NXTPAPER is the company's latest device to use NXTPAPER, TCL's custom paper-like LCD screen, which offers some of the qualities of E Ink without the limitations. TCL says the 11.5-inch color NXTPAPER Pure display on the Note A1 has a 2,200 x 1,440 resolution and a 120Hz refresh rate, which should mean it looks clearer and feels much smoother to interact with than the color E Ink screen used on something like the reMarkable Paper Pro . The tablet supports TCL's T-Pen Pro for taking notes and drawing on the screen, but also features eight built-in microphones for recording and transcribing audio. The Note A1 also has a 13-megapixel camera for scanning documents, an 8,000mAh battery and 256GB of storage, with the option to access cloud services like Google Drive or Microsoft OneDrive if you want it.

When you purchase through links in our articles, we may earn a small commission. This step-up model in TCL's Wi-Fi smart lock line includes a fingerprint reader, a keyway, and NFC support. This budget smart lock doesn't skimp on features, although reservations over the robustness of it hardware keep us from making a strong recommendation for it. TCL jumped into the smart lock market late in late 2024 with some ambitious--and pricey--locks that were hit and miss, perhaps rushed to market without enough testing. The manufacturer has since taken a step back and is fleshing out its lineup a bit more thoughtfully, including the launch of this budget-priced offering, the TCL D2 Plus Fingerprint Smart Lock Plus.

Nonlinear independent component analysis (ICA) provides an appealing framework for unsupervised feature learning, but the models proposed so far are not identifiable. Here, we first propose a new intuitive principle of unsupervised deep learning from time series which uses the nonstationary structure of the data. Our learning principle, time-contrastive learning (TCL), finds a representation which allows optimal discrimination of time segments (windows). Surprisingly, we show how TCL can be related to a nonlinear ICA model, when ICA is redefined to include temporal nonstationarities. In particular, we show that TCL combined with linear ICA estimates the nonlinear ICA model up to point-wise transformations of the sources, and this solution is unique --- thus providing the first identifiability result for nonlinear ICA which is rigorous, constructive, as well as very general.

Nonlinear independent component analysis (ICA) provides an appealing framework for unsupervised feature learning, but the models proposed so far are not identifiable.

This inexpensive smart lock covers the basics--and even provides a fingerprint reader--making for a very affordable smart lock if you don't need any other bells and whistles. You probably know TCL for its TVs, soundbars, and smart appliances more than its home security devices, but the manufacturer now offers no fewer than six smart locks (and one home security camera, too). One of its most ambitious smart locks--the D1 Pro Palm Vein Smart Lock--was a better value than the even more ambitious TCL D1 Max 3-in-1 Video Smart Lock, which boasted an integrated video doorbell. Both of those devices had their flaws, but if you're willing to give up palm vein scanning technology and an integrated camera to see your visitors, the far more basic TCL D1 is the better value. This less-expensive alternative retains the PIN pad, fingerprint scanner, and Wi-Fi connectivity in a package that's available at Amazon for just 90.

Ensuring the safety of vulnerable road users through accurate prediction of pedestrian crossing intention (PCI) plays a crucial role in the context of autonomous and assisted driving. Analyzing the set of observation video frames in ego-view has been widely used in most PCI prediction methods to forecast the cross intent. However, they struggle to capture the critical events related to pedestrian behaviour along the temporal dimension due to the high redundancy of the video frames, which results in the sub-optimal performance of PCI prediction. Our research addresses the challenge by introducing a novel approach called \underline{T}emporal-\underline{c}ontextual Event \underline{L}earning (TCL). The TCL is composed of the Temporal Merging Module (TMM), which aims to manage the redundancy by clustering the observed video frames into multiple key temporal events. Then, the Contextual Attention Block (CAB) is employed to adaptively aggregate multiple event features along with visual and non-visual data. By synthesizing the temporal feature extraction and contextual attention on the key information across the critical events, TCL can learn expressive representation for the PCI prediction. Extensive experiments are carried out on three widely adopted datasets, including PIE, JAAD-beh, and JAAD-all. The results show that TCL substantially surpasses the state-of-the-art methods. Our code can be accessed at https://github.com/dadaguailhb/TCL.

The open-world assumption in model development suggests that a model might lack sufficient information to adequately handle data that is entirely distinct or out of distribution (OOD). While deep learning methods have shown promising results in handling OOD data through generalization techniques, they often require specialized hardware that may not be accessible to all users. We present TCL, a lightweight yet effective solution that operates efficiently on standard CPU hardware. Our approach adapts contrastive learning principles specifically for tabular data structures, incorporating full matrix augmentation and simplified loss calculation. Through comprehensive experiments across 10 diverse datasets, we demonstrate that TCL outperforms existing models, including FT-Transformer and ResNet, particularly in classification tasks, while maintaining competitive performance in regression problems. TCL achieves these results with significantly reduced computational requirements, making it accessible to users with limited hardware capabilities. This study also provides practical guidance for detecting and evaluating OOD data through straightforward experiments and visualizations. Our findings show that TCL offers a promising balance between performance and efficiency in handling OOD prediction tasks, which is particularly beneficial for general machine learning practitioners working with computational constraints.

It was an interesting year for robots at CES 2025. While we had hoped the AI boom would bring a new wave of useful robots to the show, it seems that many robotics companies are still figuring out exactly how to best use AI. What we found instead was a mix of adorable robot companions, strange concepts and one, slightly terrifying humanoid. We visited a lot of robots at CES and, for better or worse, some really left an impression on us. These are the ones that stood out the most.