Mila

Evolving agent harnesses can yield a 20-35% performance boost in enterprise environments without altering model weights.

Self-supervised pre-training on just one real table can yield surprisingly strong generalization in Tabular Foundation Models, challenging conventional wisdom about dataset size.

Privileged Value Functions can inject crucial token-level signals into LLM reinforcement learning, leading to substantial performance gains over traditional methods.

Verifying the self-consistency of probabilistic AI predictions can now be achieved in polynomial time, paving the way for safer AI systems.

ERRLESS reveals that incorporating Bayesian principles into symbolic regression can yield more interpretable and accurate models, outperforming traditional methods in uncertainty quantification.

Search-oriented rubrics redefine how we evaluate document sets, leading to a 2.6-point performance boost in deep research tasks.

Explicit relational semantics can boost agent coordination in multi-agent systems, but may compromise accuracy when truth is paramount.

Foundation model agents can achieve stable cooperation in social dilemmas by inferring behavioral similarities, defying classical game theory's expectations of mutual defection.

Target variance in reinforcement learning can be drastically reduced by analytically propagating uncertainty without restrictive policy structures.

Latent guidance in the Reinformed Dreamer leads to more robust representation learning, outperforming previous methods in diverse benchmarks.

LLMs struggle to grasp unspoken beliefs, falling short of human-level understanding in recognizing and updating implicatures.

A simple yet powerful adaptation of attention for triangle meshes outperforms existing methods, achieving state-of-the-art results in geometry-processing tasks.

Language models often prioritize short-term rewards over cooperative success, revealing a critical flaw in their decision-making processes.

CCBR empowers users to shape their recommendations through text-driven interactions, outperforming traditional models while maintaining high interpretability.

Knowledge retrieval and reasoning are the main bottlenecks in KI-VQA, but a new diagnostic benchmark reveals deeper issues in visual grounding and object identification.

Adaptive temporal discounting can dramatically enhance reinforcement learning's efficiency and flexibility, outperforming traditional fixed discount methods.

Achieving high-fidelity inference in gravitational lensing, this method reveals intricate details of galaxy mass distributions that were previously obscured by noise.

SOGR1 achieves state-of-the-art performance in knowledge graph question answering while being significantly more efficient than larger models.

Merging LoRA modules with a hypernetwork enables superior few-shot adaptation, outperforming existing methods that overlook source domain knowledge.

MOJO outperforms traditional supervised learning models by effectively utilizing unlabelled data, achieving state-of-the-art results in neural decoding tasks with minimal labelled input.

WebSwarm's innovative recursive delegation allows agents to not only search but also adaptively collaborate, leading to superior performance in complex web search tasks.

Adversarial training on just 66 carefully chosen statements can achieve near-zero attack success rates, revolutionizing safety alignment in language models.

The probability of a model trajectory entering unsafe regions can be exponentially small, but the geometry of the unsafe set critically influences how quickly training stabilizes.

Fast transductive rates in semi-supervised learning can be achieved with fewer labels than previously thought, thanks to the power of data augmentation.

LLMs possess an internal estimate of their remaining output length, revealing a surprising layer of planning in their generation process.

Even advanced LLMs struggle to prevent privacy breaches in multi-user settings, exposing critical data spillage risks that current benchmarks overlook.

Retaining past knowledge can actually impede real-time adaptation in dynamic environments, leading to a new framework for optimizing continual learning.

Gemma 4's unified architecture and reasoning mode enable it to outperform larger models in human-rated tasks while maintaining high efficiency.

Trust-region optimization can dramatically enhance the training of neural quantum states, achieving stability and speed at unprecedented scales.

Gradual transitions in training objectives can significantly enhance model performance during adaptation, preserving valuable learned features.

Spectral Envelope Theory unlocks a new level of fidelity in generating synthetic relational time series, outperforming traditional methods in capturing complex temporal dynamics.

Combining hard architectural sparsity with soft regularization can significantly enhance the interpretability of vision models without sacrificing performance.

ArBG achieves a remarkable 60% reduction in zero-shot energy error for peptide systems, challenging the dominance of flow-based sampling methods.

Reusable fixing transformations can achieve a 94.3% compilable transformation rate, revolutionizing how we handle breaking API changes across multiple projects.

Allocating more capacity to earlier layers in language models can significantly enhance performance, challenging the long-held uniform layer design paradigm.

Iteratively training on a self-selected dataset can dramatically enhance vision-language model performance without the need for extra data or pre-training.

N-version programming with coding agents not only mirrors historical failures but also shows a dramatic reduction in errors through diversity, challenging assumptions about AI reliability.

FlowMaps can revolutionize robotic navigation by accurately predicting object movements based on human interaction patterns, outperforming existing methods in real-world scenarios.

Scaling self-play training directly from pixels leads to competitive autonomous driving performance without human trajectory supervision.

Discriminator-Guided RL achieves a remarkable reduction in FID from 9.38 to 2.62, showcasing a new way to align model outputs with real data without human preferences.

LSTM decoders can boost attention alignment with human fixation patterns by up to 50 percentage points, but at the cost of spatial precision and task differentiation.

Negative token filtering enables stable single-rollout training, outperforming traditional group-based methods on agentic tasks while maintaining efficiency.

MambAdapter achieves superior performance in audio and speech tasks while drastically cutting down on computational resources.

Control interventions are often detected by LLMs, with awareness levels varying significantly across models and tasks, revealing vulnerabilities in AI safety protocols.

STORM transforms lexical query expansion by turning retrieval rewards into actionable token-level signals, enabling efficient and effective query rewriting that rivals larger models.

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

Freezing parameter groups based on Fisher structural drift can enhance fine-tuning efficiency and boost performance in downstream tasks.

SHAPO redefines safe exploration by biasing policy updates toward caution in uncertain regions, leading to significant improvements in safety and performance.

No single TTS model excels across all languages, exposing the limitations of current multilingual approaches in low-resource settings.

Prefix failure in on-policy distillation can be effectively mitigated by correcting problematic prefixes, leading to significant improvements in reasoning coverage and accuracy.

Code-switched speech can exploit safety weaknesses in LALMs, achieving jailbreak success rates that challenge current safety protocols.

Counterfactuals, often seen as safe tools for model explanation, can inadvertently expose sensitive training data through membership inference attacks.

Active exploration can dramatically enhance adults' ability to reason about complex causal relationships, but even with this advantage, they still struggle compared to simpler tasks.

AuditFlow achieves over 82% accuracy in structured financial audits by leveraging a unique symbolic environment that outperforms traditional methods by nearly 15 points.

MobEvolve outperforms traditional methods by evolving its logic through targeted updates, achieving unprecedented fidelity and interpretability in human mobility generation.

Language specialization in multilingual MoEs happens mostly in the final layers, suggesting a surprisingly simple recipe for parameter-efficient adaptation.

Human-generated citation lists, long considered the gold standard for evaluating literature search, are surprisingly unreliable, with LLMs judging them relevant only ~50% of the time.

How you represent a plan matters more than which LLM you use when building robust web agents.

Bridging the scientific knowledge gap for hundreds of millions, AfriScience-MT pioneers document-level scientific machine translation for six African languages.

Offline policy optimization with a world model allows for affective music recommendation that improves user valence and arousal, even when ethical constraints preclude online experimentation.

Uniformly quantizing the entire diffusion action head of VLAs to W4A4 is not only possible, but can match or exceed FP16 performance, defying conventional wisdom and slashing memory footprint by 71%.

Exploiting approximate label symmetries can significantly improve data scaling in machine learning, even when those symmetries aren't perfect.

Causal features aren't just for understanding data; they can also align long-term incentives in strategic classification, offering a path to robustness and fairness.

Forcing VLMs to "think" visually with panoramic renderings, rather than relying on language alone, unlocks surprisingly robust spatial reasoning.

Frustrated by biased algorithms? Now, users can band together to nudge AI systems toward fairness, even without the platform's permission.

As AI agents scale and interact, the Foundation Protocol offers a coordination layer that prioritizes accountability and governance, ensuring that the future of human-AI collaboration remains open and governable.

DINOv2's representation space is so statistically well-behaved that you can train a vanilla diffusion transformer on it and beat specialized architectures with fewer parameters.

Forget expensive downstream evaluations: token-level statistics from expert-written solutions can reliably forecast LLM performance with 10,000x less compute.

Dye-filled boron nitride nanotubes aren't the optically-bright J-aggregates we hoped for, but instead form structurally heterogeneous confined ensembles with suppressed radiative rates.

LLM-powered query reformulation, a hot topic in IR, often fails to translate gains from lexical to neural retrieval, and bigger models don't always help.

LLMs struggle with structured 2D tasks when inputs are serialized into 1D, revealing a surprising performance gap compared to vision-augmented models that directly process the 2D layout.

ManifoldRank reveals that treating fairness as a taxation cost can significantly enhance the effectiveness of online fair re-ranking algorithms.

CroSearch-R1 reveals that integrating cross-lingual knowledge through a dynamic retrieval strategy can substantially enhance the performance of Retrieval-Augmented Generation systems.

AI is poised to revolutionize protein dynamics research, but key challenges remain in ensuring scalability, thermodynamic consistency, and kinetic fidelity.

LLMs re-rank documents better when you learn to route each query to the specific attention heads that matter, instead of relying on static subsets or everything at once.

Sampling plausible configurations of digital twins can reveal multiple valid parameterizations, enhancing model adaptation in complex natural systems.

Out-of-domain self-supervised pretraining on brain MRIs beats in-domain supervised learning when generalizing to real-world clinical data.

Multi-modal alignment in symbolic regression models like SNIP doesn't actually improve during optimization, suggesting current approaches are too coarse to effectively guide symbolic search.

Stop overpaying for LLM serving: intelligently routing requests to specialized pools based on token budget slashes GPU costs by up to 42% and dramatically improves reliability.

Forget hand-crafted prompts: RL can automatically unearth 36 new failure modes in VLMs that humans miss, revealing surprising blind spots in counting, spatial reasoning, and viewpoint understanding.

Merging experts in MoE LLMs can actually *improve* performance compared to pruning, offering a new path to compression that preserves capabilities.

MoEs don't always need learned routers: routing information can be embedded directly in the hidden state.

Sparse autoencoders' failure to generalize compositionally isn't due to amortized inference, but because they learn lousy dictionaries in the first place.

Mimicking human cognition, FLAIR lets dialogue models "think while listening," boosting performance without adding latency.

Ditch the cross-world counterfactuals: Sequential Transport offers a DAG-aware, optimal transport approach to causal mediation analysis, providing deterministic counterfactual mediators and fine-grained attribution.

Forget exotic attention mechanisms – MobileLLM-Flash achieves up to 1.8x faster LLM prefill on mobile CPUs by smartly pruning and adapting existing architectures for on-device use.

Tactile sensing can be efficiently injected into vision-language-action models via feature-wise linear modulation, boosting robot manipulation performance without the computational overhead of large-scale pretraining.

Forget contrastive learning: LLM2Vec-Gen learns text embeddings by representing the *response* an LLM would generate, unlocking safety and reasoning abilities for embedding tasks.

Democratized LLM pre-training is now a reality: Covenant-72B proves you can train a competitive 72B model with untrusted peers over the internet, opening the door to broader participation and reduced costs.

Diagonal SSMs, despite their empirical success, provably fail to track states of non-Abelian groups, revealing fundamental limitations in their expressive power.

One in four initial posts on a major cybercrime forum contain explicit crime-related content, revealing a surprisingly high baseline of open criminal activity.

Forget full fine-tuning: this dynamic routing strategy lets you adapt dense retrieval to new domains while using just 2% of the parameters.

Takeuchi's Information Criterion (TIC) accurately predicts DNN generalization gaps, but only when models operate near the Neural Tangent Kernel (NTK) regime.

Quantum computers could compromise nuclear power plants' safety systems and create unsolvable forensic paradoxes, with current defenses showing alarming vulnerability.

A global consensus on AI safety risks and capabilities has emerged from a panel of 100+ independent experts, representing a landmark effort in international collaboration.

Achieve state-of-the-art dynamic graph anomaly detection with limited labels by learning a robust decision boundary around normal data, outperforming methods that overfit to scarce anomalies.

Attention-based re-ranking gets a boost: ReAttn's post-hoc re-weighting tames over-concentration and lexical bias, leading to more accurate and interpretable results without extra training.

LLMs struggle to balance rational financial decisions with mimicking noisy user behavior, often overfitting to short-term market trends instead of aligning with long-term investment goals.

Coreference benchmarks may be overstating language models' NLU abilities, as even small changes to evaluation contexts reveal a failure to generalize.

Cybercriminals are actively exploring AI's potential for both enhancing existing attacks and creating novel illicit tools, but harbor significant doubts about its real-world effectiveness and impact on their operations.