Gear The 3Doodler is a handheld 3D printer that makes a great gift and it's only $40 at Amazon for Black Friday These are the best early Black Friday deals on STEM gifts for kids under $50. We may earn revenue from the products available on this page and participate in affiliate programs. Buying gifts for kids can be hard. You want to get them something creative, but it also has to be fun enough to keep their attention. Plus, you don't want their parents to hate you for it (most of the time).

Large language models (LLMs) have demonstrated remarkable capabilities in various kinds of tasks, while the billion or even trillion parameters bring storage and efficiency bottlenecks for inference. Quantization, particularly floating-point quantization, is known to be capable of speeding up LLM inference by reducing memory footprint and data movement during the inference process. For the first time, we advance the floating-point quantization exploration from integer bitwidths to non-integer bit-widths, namely AMS-Quant, to further approach the quantization sweet spot. AMS-Quant incorporates two novel techniques to put it into effect: (1) it proposes Mantissa-bit Sharing, which groups k quantized weights and lets them share the least significant mantissa bit, allowing us to further approach the minimum quantization bit-width without accuracy loss. (2) It introduces Adaptive Searching, which employs an offline optimization strategy to minimize the accuracy degradation introduced by sharing. Moreover, AMS-Quant is also prototyped as efficient CUDA Linear kernels, which translates memory savings into wall-clock latency reduction by reducing memory access. Extensive experiments on large-scale datasets and models show that AMS-Quant can quantize the model to FP-5.33-e2m3 and FP4.25-e2m2, and significantly speed up the LLM decoding over FP16 inference (2.8x and 3.2x), with negligible accuracy loss.

Reinforcement learning (RL) policies deployed in real-world environments must remain reliable under adversarial perturbations. At the same time, modern deep RL agents are heavily over-parameterized, raising costs and fragility concerns. While pruning has been shown to improve robustness in supervised learning, its role in adversarial RL remains poorly understood. We develop the first theoretical framework for certified robustness under pruning in state-adversarial Markov decision processes (SA-MDPs). For Gaussian and categorical policies with Lipschitz networks, we prove that element-wise pruning can only tighten certified robustness bounds; pruning never makes the policy less robust. Building on this, we derive a novel three-term regret decomposition that disentangles clean-task performance, pruning-induced performance loss, and robustness gains, exposing a fundamental performance--robustness frontier. Empirically, we evaluate magnitude and micro-pruning schedules on continuous-control benchmarks with strong policy-aware adversaries. Across tasks, pruning consistently uncovers reproducible ``sweet spots'' at moderate sparsity levels, where robustness improves substantially without harming - and sometimes even enhancing - clean performance. These results position pruning not merely as a compression tool but as a structural intervention for robust RL.

Despite its success, the influence of different view choices has been less studied.

The paper studies contrastive methods for self-supervised representation learning. It studies how multiple views of the same data are used for representation learning, and how the mutual information between these views matters for downstream performance. The authors propose a theory that there is a sweet spot in the amount of mutual information between two views (not too less, not too much) such that the downstream performance is highest at this point. They empirically verify this theory for two classes of views (patches, and colors). They propose a method that simply combines existing augmentations from prior work and provides gains over them.

With the NHS still struggling, a prisons crisis still teetering and Britain's borrowing costs soaring, there are few easy jobs going in Keir Starmer's cabinet at present. But even in such difficult times, the task of convincing Silicon Valley's finest to help make Britain a leader in the artificial intelligence (AI) revolution – all while one leading tech boss uses the Labour government as a regular punching bag and others ostentatiously move closer to Donald Trump – is among the most challenging. This is the mission that has fallen to Peter Kyle, the science and technology secretary, who has become an important figure in Starmer's cabinet. If balancing the concerns over online free speech, AI's impact on the climate crisis and the threat it poses to wiping out humanity are not enough, the economic headwinds Britain is now experiencing makes the launch this week of the government's AI action plan even more important. And Kyle is worried Britain could miss the boat.

The Immersed Visor aims for spatial computing's sweet spot The $1,050 device has 4K per-eye resolution and weighs less than an iPhone 16 Pro. An Austin-based startup best known for its VR and mixed reality workspace software for other companies' headsets now has hardware of its own. The Immersed Visor appears to sit somewhere between a Vision Pro Lite and Xreal Plus: a lightweight head-worn device that creates a high-resolution spatial computing environment on the cheap (well, relatively speaking). Teased to death for months, Immersed founder Renji Bijoy finally unveiled the Visor at an Austin event on Thursday. The device, a bit more than glasses but much less than a full headset, gives each eye the equivalent of a 4K OLED screen.

AMD will start selling the Radeon RX 7900 GRE (Golden Rabbit Edition) graphics card in the US, offering users a detuned version of its 7900 XT flagship for 549. For a savings of around 350 over the latter, it has performance on par with NVIDIA's RTX 4070 Super for some games at some settings, according to AMD. It offers impressive specs for that sum, including a Navi 31 XL GPU with 80 compute units (5120 stream processors), 160 AI accelerators and 16GB of GDDR6 memory. That's just a bit less than the 20GB of GDDR6, 96 compute units and 168 AI accelerators in the 7900 XT. With that, it offers 26 to 46 FP32 TFLOPS, a bit lower than the 700 XT's 32 to 51.6 FP32 TFLOPS.

In recommender systems, knowledge graph (KG) can offer critical information that is lacking in the original user-item interaction graph (IG). Recent process has explored this direction and shows that contrastive learning is a promising way to integrate both. However, we observe that existing KG-enhanced recommenders struggle in balancing between the two contrastive views of IG and KG, making them sometimes even less effective than simply applying contrastive learning on IG without using KG. In this paper, we propose a new contrastive learning framework for KG-enhanced recommendation. Specifically, to make full use of the knowledge, we construct two separate contrastive views for KG and IG, and maximize their mutual information; to ease the contrastive learning on the two views, we further fuse KG information into IG in a one-direction manner.Extensive experimental results on three real-world datasets demonstrate the effectiveness and efficiency of our method, compared to the state-of-the-art. Our code is available through the anonymous link:https://figshare.com/articles/conference_contribution/SimKGCL/22783382