Papers

Negotiation is a fundamental strategic interaction in management science, characterized by agents attempting to reach agreements while protecting private information, such as reservation costs and hidden valuations. A prevalent yet complex scenario involves a single seller negotiating concurrently with multiple buyers, each possessing heterogeneous, private budgets. In such settings, constrained by a limited number of communication turns, the seller must balance exploring the broader market to discover the highest valuation with concentrating sufficient turns on a single target buyer to secure the best possible outcome. Our analysis reveals a significant gap in standard Large Language Models (LLMs): while these models are linguistically proficient, they fail to act as effective economic decision-makers. Specifically, they exhibit a failure to explore the buyer pool, often fixating on the current highest bid rather than strategically investigating the market to discover latent high valuations. In this paper, we propose a specialized training recipe using Reinforcement Learning from Verifiable Rewards (RLVR). By anchoring the reward function to objective economic outcomes, the strategic balance between market discovery and surplus extraction emerges natively through the learning process. Our results demonstrate that the trained seller undergoes a multi-stage strategic evolution, learning to leverage price anchoring and strategic probing to identify more profitable counterparties. The agent extracts a substantially higher surplus than frontier models by both improving its persuasive bargaining skills and consistently closing deals with high-value buyers. Finally, we show that our seller strategies generalize robustly to unseen buyer negotiation styles and budget distributions.

Large reasoning models (LRMs) improve language model capabilities by generating explicit thinking traces before final answers. In factuality-oriented question answering (QA), such thinking often improves overall performance by helping the model recover relevant knowledge and refine its answers. However, we find that this benefit is not uniform at the instance level: explicit thinking can also overturn correct non-thinking answers and lead to factual drift. We refer to this failure mode as thinking-induced hallucination. To explain this phenomenon, we formulate explicit thinking in factuality QA as a thinking residual over the model's direct-answer tendency, which can either recover missing knowledge or introduce unsupported associations. Based on this formulation, we propose MARGO, \underline{M}ixed-Mode \underline{A}dvantage \underline{R}egularization for \underline{G}rounded \underline{O}ptimization, a reinforcement learning framework that uses non-thinking rollouts as same-model references in advantage estimation. By constructing mixed-mode rollout groups with both thinking and non-thinking trajectories, MARGO evaluates whether explicit thinking adds factual value beyond direct answering, thereby suppressing hallucination-prone thinking while preserving beneficial thinking behaviors. Experiments across multiple factuality-oriented QA benchmarks demonstrate that MARGO improves factual reliability over strong baselines, while evaluations on mathematical benchmarks show that it preserves general reasoning ability.

Information Operations on social media networks have been identified as a significant threat to democracy and modern society, but they are challenging and expensive to detect by humans. Existing supervised IO detection methods fail to capture the dynamic nature of evolving IO user behavior, while existing unsupervised approaches rely on oversimplified assumptions of coordination among IO users that may not exist in practice. To overcome the limitations of existing methods, we formulate IO user detection as an anomaly detection problem and propose a novel unsupervised IO user detection approach called Temporal-bEhavior-laNguage Signals for information Operation Recognition (TENSOR), which leverages multimodal data, including temporal online user behavior, such as message posting activities, and the textual content of the messages. The motivation is that IO users are typically a very small fraction of all online users and have unique temporal behavioral and language patterns. Specifically, we train a Temporal Point Process (TPP) to capture abnormal temporal behavioral patterns of IO users because they are known to behave in a coordinated manner for IO campaigns. We further introduce a novel evidence function that converts LLM responses, which are generated from user post timelines, into quantitative scores to adjust the TPP outputs for better IO user detection. Experimental results show that TENSOR outperforms the baselines on five real-world IO datasets. Code is available at https://github.com/xiuzhenzhang/TENSOR.

Accurate ranking of antibody candidates according to their binding affinity is essential for therapeutic antibody discovery. However, existing methods treat affinity comparisons independently and ignore the contextual information encoded in other labeled comparisons, limiting their ability to capture antigen-specific binding landscapes. For many target antigens, a small number of experimentally characterized affinity comparisons are often available. An important question is whether the model can exploit these existing comparisons to infer antigen-specific ranking patterns that facilitate subsequent affinity ranking. This form of learning from labeled demonstrations closely resembles the paradigm of In-Context Learning, motivating us to revisit antibody affinity ranking from an ICL perspective. To this end, we propose AbICL, an ICL framework for antigen-specific antibody affinity ranking. AbICL combines a pretrained structural encoder with a context ranking head and is trained with an episodic meta-training strategy that enables the model to leverage support demonstrations for test-time adaptation without gradient updates. Experiments on the AbRank benchmark demonstrate that AbICL consistently outperforms existing ranking baselines across almost all data splits and evaluation benchmarks. Further analysis shows that the value of contextual demonstrations depends on how well they match the target inference task, and becomes increasingly pronounced under distribution shift and fine-grained affinity discrimination. These findings highlight the potential of ICL as an effective paradigm for antigen-specific antibody affinity ranking, particularly in challenging settings where a single global ranking function is insufficient.

The limited-memory BFGS (L-BFGS) algorithm is a cornerstone of large-scale optimization due to its linear memory and computational costs. However, in ill-conditioned or non-convex landscapes, the implicit inverse Hessian approximation can suffer from an exploding condition number, leading to numerical instability and degraded convergence. To address this, we propose Two-Sided L-BFGS, a safeguarded variant that dynamically constrains the condition number of the inverse Hessian operator via a two-sided geometric envelope. Moreover, we show that Two-Sided L-BFGS preserves accumulated curvature information and maintains standard $O(mn)$ memory and per-iteration time complexities. We prove that this geometric envelope yields a uniform bound on the condition number of every inverse Hessian approximation generated by the algorithm. By tracking the algebraic evolution of the extreme eigenvalues through $m$ consecutive quasi-Newton updates starting from a scaled identity matrix, the resulting bound is expressed explicitly as a function of the memory depth, problem dimension, and envelope hyperparameters. Moreover, we show that Two-Sided L-BFGS preserves asymptotic global convergence in non-convex regimes under standard smoothness and strong Wolfe line-search assumptions, matching the theoretical guarantees of L-BFGS variants utilizing the Li-Fukushima cautious update rule. Numerical experiments on high-dimensional optimization problems demonstrate that the proposed method maintains well-conditioned inverse Hessian approximations and improves robustness and convergence behavior on ill-conditioned benchmarks.

The increasing uncertainty from flexible demand and renewable generation has made distributionally robust optimization (DRO) an important tool for robust power system dispatch. DRO relies on forecast scenarios to construct ambiguity sets, but conventional scenario generation pipelines are often trained in an accuracy-oriented manner and may neglect spatial correlations among uncertainties. This mismatch can produce ambiguity sets that are statistically plausible but suboptimal for downstream operation. This work proposes a decision-focused generative framework for correlated scenario generation in DRO-based dispatch. Instead of training generative models solely to fit the historical uncertainty distribution, the proposed framework optimizes generated scenarios according to their induced downstream operational cost. The proposed framework is tailored to mainstream generative models, including variational autoencoders, generative adversarial networks, and diffusion models, while capturing the joint distribution of uncertainties across buses. To improve computational tractability, we further develop a differentiable scenario selector that selects decision-relevant scenarios from a generated pool and can be trained within the same decision-focused pipeline. Case studies demonstrate that the proposed framework effectively reduces 0.80%-2.02% operational cost across different generative models compared to accuracy-oriented methods.

The Minkowski functionals of a field's excursion sets – area, boundary measure, and Euler characteristic – describe its level-set morphology; the Euler characteristic is the cheapest handle on topology. We derive smooth Monte-Carlo estimators for all three of a continuous neural field, evaluated at scattered points via the co-area formula and Gauss-Bonnet, using only autodiff: no grid, no complex, no persistence. The estimator is accurate to 1-3% against exact topology in 2D and 3D, and costs about 3 ms per iteration where a persistent-homology (PH) loss on a cubical grid costs 650-1000 ms – a 250x gap. We establish four design rules without which these losses silently fail: a dense level ladder (invariants are flat in the parameters away from transitions), a $C^2$ backbone (ReLU nets hide curvature in kinks), the full Minkowski vector (Euler characteristic alone is an alternating sum, gamed by debris-hole cancellation; pricing perimeter closes the channel), and sampling-scale coverage. In 2D the vector-valued cap is the only method in a controlled comparison that both repairs topology (3/3 seeds) and preserves fidelity – uniform smoothing repairs at 11-17x the fidelity cost, and the Euler term alone repairs nothing. In 3D neural-SDF fitting, however, a failure mode we believe general to any sampled soft topology objective appears: gradient descent adversarially hides topological noise below the sampling density, where the estimator is blind – spurious-feature counts are invariant to 4x more samples, and closing the window needs cubically many points, erasing the cost advantage. A grid-based PH baseline, whose complex is the evaluation resolution, solves the same benchmark ($4/9$ exact; median $b_1$ error 1 vs. ours above $10^4$). The 250x cost of persistence is, at present, the price of having no null space. We release estimators, receipts, and benchmarks.

Real-time decoding is a major bottleneck in scaling quantum error correction (QEC) from noisy intermediate-scale quantum (NISQ) devices to fault-tolerant quantum computing. We present an adaptive confidence-gated decoding framework for the rotated surface code that treats decoding as a two-stage inference problem. A lightweight feed-forward neural network performs fast-path decoding for the majority of syndrome measurements, while only low-confidence predictions are escalated to a minimum-weight perfect matching (MWPM) refinement stage. We benchmark the framework on rotated surface codes with distances $d \in \{3,5,7,9,11\}$ under circuit-level depolarising noise using the Stim stabiliser simulator. The evaluation characterises logical accuracy, confidence-controlled accuracy-latency trade-offs, decoding throughput, per-shot latency, and decoding-graph resource scaling. Routing only 3.3%-6.2% of syndromes to the refinement stage improves logical accuracy from 99.21% for the neural-only baseline to 99.81% at a confidence threshold of 0.95 while incurring only a bounded increase in average decoding cost. Neural-decoder throughput saturates near $4.6 \times 10^{5}$ samples s$^{-1}$ at batch size 512 on commodity CPU hardware, indicating that the neural fast path is not the dominant throughput bottleneck beyond code distance $d=7$. We release the complete benchmarking pipeline, trained models, raw benchmark data, and source code, and explicitly distinguish the experimentally validated contributions from the broader hardware-aware QEC co-design roadmap, including hardware-constrained code discovery, GPU-accelerated inference, and multi-noise optimisation, which remain directions for future work.

We study repeated procurement auctions in which producers have private costs and the platform must learn the context-dependent value of selecting each producer. We evaluate performance by welfare regret: the cumulative loss in total surplus relative to the full-information efficient rule that knows the context-dependent values and true producer costs. The natural UCB allocation rule achieves $\widetilde O(\sqrt{ngT})$ welfare regret under truthful bids, but its adaptive, bid-dependent learning path does not by itself ensure truthfulness. To obtain exact incentives, we first design a bid-independent explore-then-commit mechanism with empirical threshold payments; it is dominant-strategy truthful and has $\widetilde O((ng)^{1/3}T^{2/3})$ regret. We then introduce frozen-payment UCB, which estimates payments from initial bid-independent exploration but continues allocation learning by UCB. Under a truthful-path margin condition, the frozen-payment UCB is approximately truthful with an average per-round deviation gain $\widetilde O(T^{-1/4})$ for fixed $n$, $g$. Under truthful bidding, it achieves $\widetilde O(\sqrt{ngT})$ welfare regret, matching the UCB rate. A lower bound shows that this regret-incentive tradeoff is tight within the frozen critical-payment class.

Training data for machine learning is routinely collected by a selection process the model never sees: loans are observed only when granted, outcomes only when a test was ordered. The standard fixes – importance weighting, covariate-shift correction, MAR imputation – assume selection is ignorable given observables. Econometrics solved the harder case in 1979: Heckman's two-equation model jointly fits a probit selection equation and an outcome equation linked through correlated errors, and the inverse-Mills-ratio term corrects for selection on unobservables, where importance weighting is structurally helpless. We instantiate this for deep epistemic uncertainty: a deep outcome network, a linear selection head, and a joint bivariate-normal likelihood over all units, ensembled for predictive variance. In a controlled generator where sampling probability depends on an unobservable correlated (rho up to 0.9) with the outcome noise, deep ensembles, MC dropout, and GP baselines are overconfident exactly where data was avoided: coverage of nominal-90% intervals falls to 64.4% at rho=0.9, and importance weighting with oracle propensities does not fix it (43.1%) – reweighting corrects the covariate distribution, not the conditional bias E[y|x,selected] != E[y|x]. The Heckman correction restores coverage (88.9%) when the selection equation has an instrument – a variable affecting selection but not the outcome – and degrades measurably without one (40.3%); we chart this honesty curve rather than hide it. On real tabular data with induced MNAR selection, the corrected intervals are the best-calibrated (lowest region-ECE) non-oracle method in selected-against regions; baselines matching its raw coverage do so only by over-widening everywhere. Our estimators reproduce classic Stata output to seven digits. We state which identification regime a practitioner is in, and release the code.

Dilution refrigerators are the enabling infrastructure of superconducting quantum computers, yet their fault diagnosis is still dominated by threshold alarms that report that something is wrong, not what. We present Onnes, a physics-grounded digital-twin simulator of a dilution refrigerator (a forward physics model with a learned real-fridge noise fingerprint) that drives a live multi-agent LLM operations layer, and use it for a controlled head-to-head between a zero-shot LLM agent panel and a supervised ML classifier on cryogenic fault diagnosis. The twin couples a real dilution-cooling floor, a noise-and-correlation fingerprint learned from real BlueFors logs, and six physics-grounded fault classes, three engineered to overlap on temperature but separate on flow and pressure. Across a 1000-turn evaluation the zero-shot panel shows no significant difference from the classifier on detection but trails on classification, its errors concentrating on the confusable faults. Curated contrastive few-shot demonstrations and self-consistency voting then raise classification accuracy from 0.685 to 0.990, matching the supervised classifier (0.985) with no parameter updates and six labeled demonstrations; an ablation attributes the gain almost entirely to the demonstrations. Run as a continuous monitor across a nine-run fault-by-seed sweep, the agent catches every developing fault within one poll interval, and a confidence gate suppresses pre-onset false alarms whose rate is backend-dependent. As a first sim-to-real check, a detector trained purely on real BlueFors telemetry posts a real-hardware false-alarm rate of 6.4% and 100% recall on physics faults injected onto real held-out windows. All numbers are drawn verbatim from released run logs.

Boosting is a fundamental technique for generically improving the accuracy of learning algorithms (Schapire 1989). Existing boosting algorithms construct a strong learner using $O(\log(\frac{1}{\epsilon})/\gamma^2)$ calls to a $\gamma$-advantage weak learner, and this round complexity is known to be optimal for generic boosters that succeed on all concept classes (Freund 1995). We show that this lower bound can be circumvented for concept classes that satisfy a mild closure property. Specifically, we present a new boosting algorithm that, for any class $\mathcal{F}$ closed under $O(\log \frac{1}{\gamma})$-XOR, strong learns $\mathcal{F}$ using $O(\log \frac{1}{\epsilon})$ calls to a $\gamma$-advantage weak learner and a single batch of $\tilde{O}(\log(\frac{1}{\epsilon})/\gamma^2)$ additional samples. Our algorithm arises from a new and simple connection between boosting and list-decodable codes. Viewing the target function as a message, we run the weak learner on its encoding and view the resulting weak hypothesis as a corrupted codeword. Feeding this corrupted codeword to a list decoder, we obtain a small list of candidate hypotheses, at least one of which is a strong hypothesis for the original function. Using additional samples, we identify and output this strong hypothesis.

Automated experimentation is moving from closed-loop optimization toward open decision-making, where human or AI planners must forecast the consequences of candidate actions before executing them. Such forecasts require a model of both sides of the experiment: how the sample is likely to respond and what the instrument is likely to detect. We therefore introduce a coupled digital-twin framework that separates these roles and then links them. In this framework, the sample twin encodes material state inferred from prior knowledge and measurements till the moment. The instrument twin captures signal formation, feedback dynamics, and operating constraints based on prior knowledge. When coupled, the two twins estimate expected outcomes, uncertainty, and risk for candidate microscope operations. For amplitude-modulation scanning probe microscopy, we realize this framework with a physics-informed encoder of force-distance curves, a deterministic scanner model of cantilever and feedback dynamics, and sparse learned residual corrections. The encoder first recovers scanner-driving descriptors with sub-nanometer accuracy. The calibrated scanner then reproduces typical traces within a few nanometers and identifies operating-point noise amplification as the main source of mismatch. Supplementary phase analysis localizes residual error to the phase channel, which clarifies where added physics is needed. Together, these results establish coupled sample and instrument twins as a practical foundation for predictive microscope operation and autonomous experimental planning.

The community has recently developed various training-time defenses to counter neural backdoors introduced through data poisoning. In light of the observation that a model learns poisonous samples responsible for the backdoor easier than benign samples, these approaches either use a fixed threshold of the training loss for splitting or iteratively learn a reference model as an oracle for identifying benign samples. In particular, the latter has proven effective for anti-backdoor learning. Our method, HARVEY, leverages a similar yet crucially different technique: learning an oracle for poisonous rather than benign samples. Learning a backdoored reference model is significantly easier than learning a reference model on benign data. Consequently, we can identify poisonous samples much more accurately than related work identifies benign samples. This crucial difference enables near-perfect backdoor removal as we demonstrate in our evaluation. HARVEY substantially outperforms related approaches across attack types, datasets, and architectures, lowering the attack success rate to the very minimum at a negligible loss in natural accuracy. The figure below shows an overview of our methods working principle.

Molecular property prediction often relies on isolated data modalities, where continuous 3D graph neural networks (GNNs) struggle to efficiently capture long-range topological dependencies and exact macroscopic heuristics. In this work, we introduce a parameter-efficient Tri-Branch Modular Fusion Neural Network that synthesizes three orthogonal modalities: 3D spatial geometry (SchNet), discrete topological grammar (SMILES via ChemBERTa), and explicit macroscopic physicochemical descriptors (Deep & Cross Network). By bypassing standard scalar readouts and employing a shared late-fusion architecture, the framework establishes a mathematically rigorous multimodal latent space that effectively resolves the arithmetic and oversmoothing limitations of local message passing. We evaluate the proposed architecture on the QM9 benchmark, targeting the extensive thermodynamic property of atomization energy at 0 K ($U_0^{\mathrm{atom}}$). Through systematic combinatorial ablation and latent bottleneck optimization ($d_e=64$), the tri-modal framework achieves a validation Mean Absolute Error (MAE) of 0.0207 eV. Operating with fewer than one million parameters, this architecture decisively surpasses the sub-chemical accuracy threshold and yields a substantial 20.6% error reduction over a strictly controlled geometric baseline. Ultimately, our findings demonstrate that integrating orthogonal macroscopic and topological data streams provides a synergistic, $\mathcal{O}(1)$ physical shortcut. This multimodal alignment offers a highly efficient alternative to brute-force parameter scaling, establishing a robust surrogate model for high-throughput virtual screening (HTVS) pipelines.

Infinite-width limits are a standard way to reason about neural networks, but it is not automatic that the limiting learner has the same complexity-theoretic inductive bias as large finite networks. We study this question for Bayesian neural networks at the mean-field, or critical feature-learning, scaling. The central quantity is the reduced entropy \[ s_\infty(y,\varepsilon)=\limsup_N -\frac{1}{N}\log \pi_N^0(L\le \varepsilon), \] the intensive prior cost of representing a target function $y$ to population mean-squared error $\varepsilon$. Our main result is a width-robust learnability theorem. At fixed depth, a family of Boolean-cube targets is learnable from polynomially many samples at infinite width if and only if it is learnable at polynomial width, if and only if its reduced entropy is polynomially bounded. Equivalently, up to polynomial slack in accuracy, the Bayesian mean-field learner generalizes exactly on the targets that can be represented by polynomial-size networks. The forward direction is proved by a form of subsampling: from the infinitely many hidden neurons in the mean-field solution, one can select polynomially many representatives and still preserve the learned function on every input simultaneously. At the critical scaling this subsampling has both an “active” component, which keeps the data-dependent low-dimensional statistics, and a “lazy” component, which resamples the entropy-dominated directions from the prior. Thus the infinite-width mean-field limit gives a clean analytic description of learning without introducing spurious width-dependent generalization power.

Association unlearning aims to disable learned label-attribute shortcuts while preserving task performance. Existing evaluations mainly measure output-level robustness or probe whether shortcut attributes remain readable in frozen features, but neither test determines whether a retained association remains functionally usable by the original classifier. We propose the Association Restoration Test (ART), a post-hoc diagnostic for functional shortcut restorability. ART estimates class-conditional association directions, amplifies residual components, and evaluates the modified features with the original classifier head. Across Waterbirds, CelebA, SpuCoDogs, and an ISIC timestamp-artifact extension, we show that output metrics, representation probes, and ART characterize distinct aspects of shortcut mitigation. These findings motivate restoration-aware evaluation for unlearning and shortcut-mitigation methods that target learned associations rather than individual classes or concepts.

Quantum convolutional neural networks (QCNNs) combine the power of quantum computing and classical CNN for computational speedup in classification tasks. However, noise levels on state-of-the-art quantum devices remain too high for practical QCNN execution. In addition, despite the reliable surface code providing a method for error rates below a threshold value, they have a prohibitively large qubit cost. Recently introduced bivariate bicycle (BB) codes are of particular interest for their high error threshold, constant encoding rate, and linear code distance. Through simulation with realistic hardware noise sources, we demonstrate that a 4-qubit unprotected QCNN fails to converge and exhibits a worse learning rate compared to numerical simulations. Addressing both limitations, we propose a distance-4 BB quantum error-correction (QEC) technique for QCNNs. In doing so, we validate that our low-overhead QEC technique for QCNNS represents a step toward practical QCNNs.

Diffusion models have become a dominant paradigm for high-quality generative modeling, while post-training is essential for adapting them to diverse downstream applications. However, post-training of large diffusion models is still challenging due to the prohibitive memory footprints and slow training speed, which existing parameter-efficient fine-tuning methods only partially address. To overcome these limitations, we propose FourTune, an efficient post-training framework for diffusion models based on an end-to-end W4A4G4 paradigm. FourTune introduces a triple-branch hybrid pipeline that augments the standard LoRA architecture with a frozen numerical stabilizer to isolate quantization-sensitive outliers, enabling stable training under native 4-bit computation. In addition, FourTune employs hardware-efficient block-wise quantization and customized fused kernels to support efficient quantized backpropagation and reduce memory bandwidth overhead. Across customization, reinforcement learning, and distillation tasks, FourTune matches the quality of full-precision fine-tuning. On FLUX.1-dev (12B), FourTune reduces memory overhead by 2.25$\times$ and increases end-to-end training throughput by 2.27$\times$ compared to BF16 LoRA.

Peer-to-peer energy trading among electric vehicles (EVs) can improve charging flexibility under limited charging infrastructure, but effective EV–EV trading requires coordinated provider–consumer matching under journey-specific conditions. This paper proposes EVTradeMatch, a prediction-guided multi-objective optimization framework for mobility-aware EV–EV energy trading. Building on the EVNextTrade study, a prior learning-to-rank model for charging-node recommendation, we define charging-node suitability as a prediction-derived score reflecting the appropriateness of assigning a provider–consumer pair to a candidate charging node based on mobility, energy, and contextual trading features. This score is used as a guidance signal and as an explicit optimization objective rather than as a hard selection rule. The EV–EV matching problem is formulated as a multi-objective mixed-integer linear program that maximizes matching coverage, transferred energy, and charging-node suitability while minimizing mobility cost under spatial, temporal, one-to-one matching, and charging-node capacity constraints. To approximate Pareto-efficient solutions in wide-area dynamic settings, we develop a tailored non-dominated sorting genetic algorithm II (NSGA-II). Experimental results show that EVTradeMatch improves transferred energy by 74.2–82.8% and charging-node suitability by 8.3–84.4% compared with proximity- and auction-based state-of-the-art methods, while improving matching coverage by 3.74–25.07 percentage points. Balanced NSGA-II solutions achieve 53.07$\pm$0.76% matching coverage and transfer 1701.94$\pm$17.03 kWh, with higher mobility cost as an explicit trade-off against travel-minimizing methods. Pareto-front analysis shows that the framework supports flexible selection among high-coverage, high-energy, low-mobility-cost, and high-suitability solutions according to operational priorities.

Sycophancy in large language models, the tendency to prioritize user agreement over truthful responses, has been documented extensively but studied primarily in single-turn settings. This paper investigates a critical question: does subjecting LLMs to greater interaction scaffolding make sycophancy better or worse? Across 4,800 veracity judgments (200 statements $\times$ 6 models $\times$ 4 conditions), we find that the interaction scaffolding characteristic of agentic systems (feedback loops, reconsideration checkpoints, and iterative refinement) systematically amplifies sycophantic behavior. Multi-turn interaction, user pressure, and iterative self-refinement each provide additional opportunities for models to drift toward agreement, and this drift coincides with a mean accuracy drop of $-6.3$ percentage points, establishing the capitulation as harmful rather than corrective. More capable models show larger amplification effects, a troubling inversion of expectations. We introduce the concept of agentic sycophancy amplification (ASA) and two novel metrics: capitulation rate and sycophantic capitulation rate. Our results indicate that as AI systems acquire greater autonomy, sycophancy becomes compounding rather than merely persistent. Systems designed with human oversight loops may inadvertently create the conditions for this drift.

Transformer architectures rely on dense self-attention to model long-range dependencies, but this mechanism exhibits quadratic complexity with respect to sequence length. We introduce BCMT (Blockwise Causal Memory Transformer), an architecture for long-context language modeling that decouples local token interactions from global context propagation. Dense causal self-attention is applied independently within local blocks, while each block produces an adaptive summary aggregated through an exponential causal memory. This memory is subsequently injected back into the token representations, enabling efficient propagation of long-range contextual information without relying on explicit global attention. Unlike standard Transformers and recurrent memory architectures, BCMT maintains neither dense interactions between distant tokens nor learned memory states. Its memory mechanism is fully parallelizable and remains compatible with standard implementations of dense self-attention. Experiments on language modeling with context lengths of up to 1024 tokens show that BCMT achieves validation performance comparable to that of Dense Transformers while significantly improving training throughput and reducing memory consumption. An ablation study further confirms that these improvements arise from the proposed memory mechanism. These results demonstrate that an exponential causal memory constructed from block summaries provides an effective alternative to dense global attention mechanisms for long-context language modeling.

This position paper argues that the dominant paradigm of AI evaluation (which focuses on superhuman autonomous performance and so implicitly targets the goal of replacing humans) is guiding AI development in the wrong direction. Instead, the AI community should pivot to evaluating the performance of human–AI teams. We argue that this collaborative shift will foster AI systems that act as true complements to human capabilities and therefore lead to far better societal outcomes than will the current process.

Brain-computer interface (BCI) research relies on multistage computational pipelines, yet progress remains constrained by fragmented data formats, heterogeneous decoder implementations and hardware-specific deployment toolchains, and researchers lack an integrated workflow. Here, we fill this gap with BCIJelly, a unified computational ecosystem that integrates 18 curated BCI datasets, 15 benchmark decoders and an algorithmic library of 80 reusable modules, an automated architecture search (AAS) procedure, and hardware-aware deployment through the toChip pipeline within a single Python framework. AAS constructs task-specific decoders without manual architecture design. It is further extended into a closed-loop mode guided by a large language model (LLM), which uses task specifications, module descriptions and search history to support multitask and cross-species decoding. The toChip pipeline compiles trained decoders for execution on neuromorphic chips, enabling energy-efficient deployment for BCI systems. An accompanying visualization software provides a graphical interface to the full workflow, making BCIJelly accessible without programming. We validate BCIJelly across five BCI paradigms (motor, visual, speech, emotion and auditory) with recordings from humans, macaques and mice, and single-task, multitask and cross-species decoding settings. BCIJelly establishes a unified and extensible infrastructure that bridges decoder development and hardware-aware deployment for BCI research.

European countries are debating policies to mitigate the increased energy costs caused by renewed geopolitical tensions, while pursuing decarbonization and electrification. A notable example is Italy's 2026 Decreto Bollette package, which proposes to remove the carbon price equivalent from the bids of certain gas-driven power plants to wholesale electricity markets, among other provisions. We use this as a case study to assess the long-term implications of suppressing the carbon price signal in the electricity market for investment, emissions, and consumer costs. We employ a stylized Italian power system using MARLEY, a multi-agent reinforcement learning framework focused on long-term electricity market assessments. In this framework, we test this policy across configurations with varying levels of support for green investment, resource adequacy, and flexibility. Results show that partial suppression of the carbon price signal yields short-term cost reductions but only a minor long-term effect on total system costs, as the deferred emissions are ultimately repaid by consumers. CO$_2$ emissions rise across most configurations since suppressing the price signal erodes incentives for renewable and storage investment. Only the most ambitious configurations for supporting green investment avoid this outcome, but they do so by marginalizing the wholesale price signal itself, thereby requiring a commitment to a hybrid market paradigm that is in contradiction with the rationale of the proposed price intervention.

AI Engram (Kwon et al., 2026) formalizes the four engram criteria of neuroscience as a constrained inverse problem in weight space and solves it closed-form: concept-specific memory traces become linear objects that can be extracted once and combined arithmetically. Appendix F states the Compositional Memory States Hypothesis: edited models live on "a commutative manifold where the integration of A and B reaches a consistent equilibrium regardless of the learning sequence." The evidence base is single and paired edits – in materials terms, single-cycle tests, in which fatigue accumulation is structurally invisible. Whether the hypothesis holds under sequential load is exactly the "temporal dynamics" question the paper defers to future work. We run that test on the authors' own reference implementation, at their reported best edit strength (TOFU alpha=0.6, a choice favoring the linearity hypothesis), with pre-registered predictions, across three model charges (two vendors, two architecture families). Four findings replicate across all three: (1) zero-shot composition and sequential re-calibrated editing diverge by 61-71% of the edit magnitude; (2) cut order is not interchangeable, and the effect scales with concept overlap – in one charge the order of cutting two Paris landmarks decides whether an uninvolved third concept survives; (3) the survivors' layer-input covariances – the method's own sufficient statistics, read as strain gauges – drift monotonically with every further cut, in every surviving concept, in every charge; (4) erased knowledge partially returns under subsequent unrelated cuts. Appendix F's commutative-manifold hypothesis is thereby falsified for sequential editing; the single-edit results of the original paper are untouched. For unlearning-as-compliance: erasure certified today does not certify the artifact after its next edit.

Recommendation systems have undergone significant transformations in the past years. The transition from traditional feature interaction modules to generative next-action prediction has pushed the boundaries of personalized content. Developments have largely evolved along two separate tracks. Sequence modeling approaches on the one hand and feature interaction methods on the other. In this paper, we introduce Bumblebee, a recommendation architecture that addresses the lack of interaction between the two directions through an interleaved, stackable block design. Each block implements a micro-pipeline of layers combining sequence personalization, attention-based encoding, and feature crossing into a self-contained unit. Every block produces a joint representation of both feature modalities which is consumed by the next block in the sequence. This mechanism encourages early and repeated mixture of modalities and enriches downstream features with additional contextual information. Residual connections between blocks create cross-modal information pathways and yield additional predictive performance without adding additional parameters. Blocks can be specialized by selectively dropping components, enabling flexible trade-offs between quality and throughput. We evaluate our approach on large-scale industrial data and show consistent improvements over comparable baseline models across several classification and regression tasks. Furthermore, we conduct ablation studies to confirm that the interleaved composition itself is the primary driver of these improvements. Our results suggest that interleaving heterogeneous functional units, rather than composing deep stacks, is a promising paradigm for future-generation recommendation architectures.

The Muon optimizer reaches the grokking threshold on modular arithmetic faster than AdamW. Prior work attributes this to "spectral-norm constraints plus orthogonalized momentum" but does not isolate which mechanism matters. To better understand Moun's behavior, we run multi-seed and multi-learning-rate sweeps to decompose and stress-test the effect. First, an ablation shows the speedup comes from orthogonalization (the Newton-Schulz iteration): orthogonalize-only matches full Muon, whereas spectral-only is no faster than AdamW and is unreliable, and this verdict holds across learning rates. Second, a mechanistic analysis finds that orthogonalizing optimizers reach generalization at roughly 3x lower spectral norm and, controlling for how much the embedding actually moves, settle into a lower-norm solution rather than simply perturbing the embedding less. Third, reducing the Newton-Schulz iteration count from five to one accelerates reaching the threshold but makes the grokked solution fragile, prone to transient collapse, with fragility that grows with learning rate; a single iteration is fast and stable only at small learning rate, while the canonical five iterations are the learning-rate-robust choice. We also show spectral scaling can be dropped at no measured cost. A methodological thread runs throughout: under a stability-aware metric, "faster" claims about grokking optimizers can invert, so we report both first-crossing and sustained-grok times. To support reproducibility, we release our full training and analysis code at https://github.com/louiswang524/muon-grokking-frontier

This paper provides a rigorous, self-contained investigation of consensus segmentation. We derive the mathematical foundations from first principles – the generative model, EM algorithm, Van Leemput's marginalization analysis, identifiability conditions, Spatial STAPLE, and deep variational formulations – and validate each theoretical prediction through controlled experiments. The central finding is sobering: under common conditions, STAPLE reduces to thresholded majority voting, suffers 95% EM suboptimality, and collapses under class imbalance. These are not edge cases but typical scenarios in medical imaging. Majority voting – simple, non-parametric, and robust – is a surprisingly strong baseline that the field has perhaps too hastily dismissed in favor of more "sophisticated" methods. At the same time, the deep consensus model demonstrates that the consensus problem is not inherently difficult – it becomes tractable when the image is used alongside the labels. And conformal prediction shows that formal uncertainty guarantees are achievable and practical. We hope this work encourages practitioners to critically evaluate their consensus methods rather than applying STAPLE by default, and provides the mathematical and empirical foundation for more principled approaches.

Pre-trained foundation models (FMs) have begun transforming single-cell genomics, but scaling them raises privacy concerns. Moreover, unlike text data, single-cell data is unordered and exhibits a unique tabular structure that current single-cell FMs overlook. We introduce Tabula, a privacy-preserving FM designed with federated learning (FL) that explicitly models the tabular structure of single-cell data. To deploy Tabula, we further developed Chiron, a decentralized AI agent-enabled platform for collaborative training across institutions without sharing raw data. Beyond strong performance across downstream benchmarks, Tabula reveals combinatorial regulatory logic across diverse biological systems, including hematopoiesis, pancreatic endogenesis, neurogenesis, and cardiogenesis. Using a new scRNA-seq dataset of paired young and aged human fibroblasts, Tabula nominates rejuvenation factors through age- and identity score-guided in silico prioritization, outperforming conventional approaches. Thus, Tabula represents an important advance in single-cell foundation modeling by integrating tabular learning with FL, paving the way toward privacy-preserving virtual cells for human health.