ETH Zurich

An inexpensive time-of-flight camera can achieve reliable stabilization of an inverted pendulum, challenging the assumption that high-resolution sensors are necessary for precise control.

HDSL achieves a remarkable reduction in editing token usage by over 5 times while maintaining scene integrity and enhancing generation speed.

Systematic gaps in AI evaluation reporting are exposed, revealing inconsistencies that hinder reliable comparisons across thousands of models and benchmarks.

Evasive steganographic payloads in LLMs can be detected again by strategically recontextualizing the data, even after successful evasion of traditional methods.

Transforming acoustic identities in real-time could revolutionize applications in stealth technology and immersive sound environments.

A novel Bayesian approach reveals how to extract high-level causal concepts from diverse environments, shedding light on the intricate dynamics of cultural and political influences.

RL-trained models can significantly improve unseen language translation by effectively leveraging contextual linguistic knowledge, outperforming traditional methods.

Users are oversharing sensitive identity information at alarming rates, with a staggering 20% revealing their official ID to news websites.

NVSHMEM's innovative device-side symmetric-memory model could redefine GPU communication strategies, pushing the boundaries of hardware performance.

Agentic architectures boost network configuration repair efficacy by 12% and safety by 17%, tackling a critical source of Internet outages.

Anchor PCA reveals that focusing on shared variation across domains can significantly enhance the robustness of unsupervised dimension reduction, outperforming traditional pooling methods.

Achieving six times fewer Gaussians while surpassing state-of-the-art performance redefines efficiency in 3D scene reconstruction.

By reconstructing 3D sensor layouts from UWB distances, Ultra Diffusion Poser achieves a 22% reduction in joint position error, revolutionizing motion tracking accuracy.

Achieving 99.97% FPU utilization with O-POPE redefines efficiency in high-frequency GEMM operations, pushing the boundaries of performance and energy consumption in ML hardware.

Achieving 3.1 TOPS/W energy efficiency, Chimera sets a new benchmark for ultra-low-power AI inference at the edge.

Frequency-Weighted Neural Kalman Filters achieve up to 10% reduction in localization error by effectively suppressing noise in critical frequency bands.

Combining Gaussian noise and bilateral filtering can yield supralinear adversarial robustness in CNNs with minimal computational overhead.

Forget scaling laws: a single looped transformer block, iterated explicitly, crushes billion-parameter feed-forward networks at multi-view 3D reconstruction.

LLM-powered honeypots can trick even frontier models into longer interactions than rule-based systems, all while costing less to run.

VLMs don't lack visual understanding of quantity, they just can't connect what they see to symbolic number representations, revealing a fractured magnitude space.

Kernel methods can substantially improve off-policy evaluation for insurance pricing, enabling neural networks to discover better pricing strategies.

Over-reliance on agentic decomposition can actually *hurt* audio understanding when a strong audio frontend already provides sufficient information, highlighting the importance of conditional evidence acquisition.

Tactile sim-to-real just got real: a physics-grounded representation unlocks zero-shot transfer for complex dexterous manipulation tasks, even without ground truth sensor calibration.

Current speech translation evaluation metrics are blind to critical speech-specific information, even when given the audio signal.

Diffusion models can now drive safer and more efficient robot navigation among humans by unifying motion tracking, forecasting, and control in a single latent space.

Transformers beat LSTMs at predicting volatile stock price movements on earnings announcement days, and news sentiment consistently improves performance.

Classical SfM can get stuck, and feedforward reconstruction can be brittle, but combining them creates a system that's both robust and accurate.

Finally, a feed-forward 3D reconstruction method that spits out meshes ready for physics engines, no expensive post-processing needed.

SAC can now rival PPO in massively parallel legged robot training, unlocking its potential for efficient sim-to-real transfer and online adaptation.

Achieving robust manipulation of deformable linear objects could revolutionize automation in household tasks and manufacturing by overcoming significant challenges in trajectory accuracy.

Humanoid robots can now reliably navigate dynamic environments thanks to a new memory system that filters out gait-induced perceptual noise and focuses on persistent environmental changes.

Forget about messy concept soups – SeqLoRA lets you teach your diffusion model 100+ new tricks without them blurring together.

Complex sequence models don't always outperform simpler tabular models for predicting antibiotic stewardship interventions in pediatric ICUs, and can even suffer from worse calibration.

Finally, a feed-forward method cracks dynamic 3D scene reconstruction from multi-view video without needing camera poses, opening the door to real-time applications.

Escape the greedy trap: Convex optimization yields tokenizers that compress better and come with optimality guarantees.

Political ideology prediction gets a boost: injecting LLMs with knowledge graphs of MP relationships significantly improves accuracy.

Granular Mixture-of-Experts can now be efficient: AIR-MoE's two-stage routing slashes routing costs without sacrificing performance.

Production VLMs like GPT-4, Claude Opus, Gemini, and Grok can be easily manipulated into confidently providing false information via subtle adversarial perturbations to images, even without compromising model alignment.

Despite the promise of multimodal context, current audio-language models struggle to leverage clinical information for dysarthric speech recognition, even degrading performance in some cases.

Training a 1024-node SOM on a billion-sample dataset in just over 6 minutes shatters previous scalability limits, thanks to a novel framework that leverages multi-GPU execution, out-of-memory streaming, and flexible topologies.

Deterministic decoding can outperform stochastic self-consistency in constrained domains by systematically exploring high-probability reasoning traces, leading to better performance with less computation.

AI-driven summaries of public consultations can systematically exclude dissenting voices, raising concerns about biased policy recommendations even when individual outputs seem reasonable.

TurboQuant's claimed advantages over RaBitQ in quantization don't hold up under rigorous, reproducible comparison, raising questions about its practical utility.

Forget heavyweight processes and bandwidth bottlenecks: Proxics offers a lightweight programming model that unlocks the potential of near-data processing with efficient virtual processors and optimized communication channels.

Early layers of language models capture human-like processing signatures in reading, rivaling traditional measures like surprisal in predicting initial eye movements.

Neural operators can achieve uniform convergence rates for approximating solution maps across diverse geometric domains, challenging traditional assumptions about shape-dependent PDE solutions.

Trustworthy super-resolution in surgery is now achievable, with a model-agnostic method that identifies and mitigates unreliable reconstructions in real-time.

Bridging the gap between trust region methods and PPO, this new framework guarantees performance improvements while outperforming existing algorithms in stability and effectiveness.

GSQ closes the accuracy gap in low-precision quantization, achieving results comparable to complex vector methods while remaining easy to implement.

SCENIC delivers the best of both worlds: the high bandwidth and software integration of commercial SmartNICs, plus the customization and data processing offload capabilities of research prototypes.

RadAgent doesn't just give you the answer; it shows its work, offering clinicians a transparent, step-by-step reasoning trace for AI-generated CT reports.

Blind predictions of cyclobutanone photochemistry reveal that nonadiabatic molecular dynamics can qualitatively capture experimental results, but the accuracy of underlying electronic structure calculations remains a key bottleneck.

LLMs struggle to simulate culturally nuanced emotional responses to bureaucratic processes, especially in Eastern cultures, suggesting current models lack the socio-cultural understanding needed for accurate policy simulation.

Supercomputers can evolve beyond just pre-training to become comprehensive "AI Factories" by adopting hybrid cloud-native architectures that support the entire lifecycle of foundation models.

Europe's collaborative EPAC chip delivers a heterogeneous RISC-V accelerator, showcasing a path towards domain-specific HPC hardware built with open standards.

Continual learning just got a turbo boost: C-Flat Turbo cuts training time by up to 25% without sacrificing accuracy, thanks to a clever gradient-skipping trick.

Unlock up to 25.7% accuracy gains on frozen LLMs in knowledge-intensive domains, without any retraining, by dynamically rewarding reasoning steps.

Achieve perceptually superior video compression at extremely low bitrates by using implicit neural representations to condition diffusion models, outperforming even VVC and prior neural codecs.

Self-supervised learning on heterogeneous neutrino detector data enables foundation-style models that achieve state-of-the-art performance with an order of magnitude less labeled data.

Training 3D avatar diffusion models on millions of in-the-wild videos is now possible, thanks to a clever 3D tokenization and visibility-aware training strategy that overcomes partial observability.

Polarization isn't always about echo chambers: Europeans can agree on *what* happened in the Ukraine war, but vehemently disagree on *why* it matters.

Unlock interactive digital twins from messy, real-world videos: FunRec automatically turns egocentric RGB-D recordings into simulation-ready 3D scenes.

Scaling up avatar pre-training to 1M in-the-wild videos unlocks emergent generalization capabilities like relightability and garment support, even without direct supervision.

AI agents are far better at automating data engineering tasks than previously thought, but flawed benchmarks are obscuring their true potential.

Forget brittle, overfit skills – Trace2Skill distills diverse execution experiences into transferable agent skills that boost performance by up to 57.65% on unseen tasks, even when transferring skills learned by smaller models to larger ones.

Forget about retraining: MUNKEY offers zero-shot machine unlearning by simply deleting instance-identifying keys, outperforming traditional post-hoc methods.

Current computational aberration correction methods struggle to generalize across different camera lenses, but this new benchmark and analysis pinpoint the key factors holding them back.

Unlock superior trajectories in complex environments with a new ADMM-based solver that jointly optimizes spatial and temporal domains, eliminating the need for complex warm starting.

Ditch the slow diffusion grind: Marigold-SSD delivers zero-shot depth completion in a single step, rivaling discriminative models in speed while retaining diffusion's accuracy.

Get 6x the RLHF alignment for your LLM with a new active learning pipeline that focuses on annotating the most informative response pairs.

Clever reticle placement on wafer-scale systems can boost throughput by 2.5x and slash latency by over a third, offering a hardware-level speedup for LLM training.

LLMs can follow detailed code refactoring instructions, but still fall short of mimicking human refactoring choices in real-world codebases, highlighting a critical gap in their ability to autonomously improve code quality.

Finally, a virtual try-on system that actually works: Gaussian Wardrobe lets you swap clothes between 3D avatars with high-fidelity garment dynamics by learning shape-agnostic garment layers.

Ditch the memory banks and prototype comparisons: this method learns a compact, parametric model of normal image embeddings with an autoregressive CNN, slashing inference time and memory in unsupervised anomaly detection.

Multimodal models often exhibit lower confidence than their unimodal counterparts when they're about to fail, and this work leverages that insight to build a better failure detector.

Reasoning can boost LLM opinion alignment, but it's not a silver bullet for removing bias in political digital twins.

Forget computationally expensive fluid dynamics: this work shows that a simple, stateless model, carefully calibrated to real-world data, can create surprisingly effective digital twins for soft underwater robots.

LLM benchmark translations can be dramatically improved by test-time compute scaling, revealing a surprisingly cheap way to get more reliable multilingual evaluations.

Forget solo Git tutorials—GitAcademy's split-screen view, mirroring a partner's actions in real-time, makes learning collaborative workflows feel less like a lonely commit and more like a team sport.

Forget temperature scaling: JUCAL calibrates aleatoric and epistemic uncertainty in classifier ensembles, achieving SOTA results with significantly smaller ensembles and lower inference costs.

Unlock domain generalization with unlabeled data by exploiting the structure of anti-causal relationships, where outcomes cause covariates.

E-graphs, typically confined to isolated optimization steps, can now persist as a first-class citizen within the compiler's intermediate representation, unlocking broader and more flexible program optimization.

Forget complex architectures: RaCo achieves SOTA keypoint matching and repeatability by cleverly combining ranking and covariance estimation in a lightweight network, trained without covisible image pairs.

Achieve >97.5% of full-data VIT performance with only 16% of the data using ScalSelect, a surprisingly effective and scalable training-free data selection method.

Context files like AGENTS.md, intended to guide coding agents, often *hurt* performance and increase costs, challenging the common practice of using them.

To address the ethical imperative of improving access equality for CGT in Europe, further policy reforms are proposed including concurrent HTA, early benefit assessment, and incorporation of additional elements of value in HTA evaluations, alongside current initiatives to increase cross-border collaboration.

An interpretable deep learning model, ECG-XPLAIM, rivals ResNet in arrhythmia detection sensitivity while offering crucial insights into its decision-making process via Grad-CAM.

Multimodal LLMs often perform worse with more modalities because they struggle to jointly recognize and reason across modalities, a problem solvable with simple prompting strategies.

A new deep learning model slashes the error rate for BMI estimation from smartphone photos, opening the door to more accessible and convenient health assessments.

Automating CAD design from text prompts is now feasible, with visual feedback loops boosting performance, especially for multimodal LLMs.

Achieve faster, near-optimal path planning in complex 3D environments by combining any-angle search with multi-resolution grids, outperforming even sampling-based methods.

LLMs that excel at math don't necessarily make good math tutors, revealing a surprising trade-off between subject matter expertise and pedagogical skill.