Meta's new terms appear to give it permission to crawl and scrape websites "associated" with the businesses that use its platforms - including online publishers…
Counterfeits of luxury goods are a major problem for fashion houses. Counterfeiters are getting better at replicating brands’ marks and passing off dupes, known as…
Release v5.20.1 Special surprise including the new model Beam from Reflection AI :partying_face: Beam is a token-efficient, frontier-scale model focused on coding and agentic use cases. It is built for teams that want to run a powerful model on their own hardware, infrastructure, or cloud, and customize it to their specific needs. Links: [Documentation](https://huggingface.co/docs/transformers/main/en/model_doc/beam) Fix small bu
Replit CEO Amjad Masad envisions a future of “self-driving companies,” where AI agents handle day-to-day tasks and people focus on higher-level decisions.
Large language models (LLMs) are increasingly used to automate AI research, including language-model (LM) development, offering a potential path toward recursive self-improvement. However, turning experimental feedback into a learned ability to propose effective improvements across models faces two challenges. First, training such an ability with real feedback is expensive: reinforcement learning (RL) requires evaluating tens of thousands of proposals, each potentially requiring a full model-training run. Even at ten GPU-hours per run, reward collection alone would consume hundreds of thousands of GPU-hours, motivating a cheaper surrogate. Second, surrogate feedback can become unreliable as policy optimization favors overestimated proposals when the proposal distribution shifts. To address these challenges, we propose LM2LM, which trains LLMs with RL to generate effective LM improvement proposals. To provide scalable feedback, we use a research world model (RWM), a frozen LLM that predicts proposal gains from real experimental records provided in context, as the reward model for RL. To maintain feedback reliability, we periodically execute current-policy proposals ranked highly by the RWM and add their measured outcomes to its context. Across eight autoregressive and diffusion LM development environments, real-execution evaluations on held-out tasks show that LM2LM enables Qwen3.8-27B to outperform GPT-6-sol, even when GPT-6-sol is given 300 real experimental records in context, while requiring 33% lower data-generation cost. LM2LM also generalizes to a model family unseen during training.
Low-rank adaptation (LoRA) has become the standard method for parameter-efficient fine-tuning of large pretrained models, yet theoretical explanations for why low-rank updates suffice remain incomplete. Most existing analyses rely on the Neural Tangent Kernel (NTK) regime, which linearizes the network around the pretrained weights. In this work, we take a different approach that requires no linearization: our key observation is that, once the backbone is frozen, the loss sees each adapted weight matrix only through a finite-rank linear map induced by the frozen activations. Trace-norm regularization then forces the optimizer to pick a low-rank matrix among all matrices that produce the same observed output. This yields global minimizers whose ranks are controlled by the rank of this map, which we call the bottleneck rank, rather than the sample size. For multi-head attention, we obtain tighter rank upper bounds by exploiting two additional invariances: row-wise softmax is insensitive to additive row constants in the score matrix, and value updates are observable only after attention weighting and frozen output projection. For the factorized LoRA objective with weight decay, we prove a complementary result on the nonconvex side: after an arbitrarily small generic positive semidefinite perturbation, every first-order stationary point is rank-deficient once the LoRA rank exceeds a threshold determined by the bottleneck rank. Finally, we show that global minimizers of the perturbed objective are near-optimal for the original trace-norm problem, and we provide uniform generalization bounds governed by the spectral structure of the frozen features.
Tool-augmented language-model agents rely on interface text: tool descriptions, argument schemas, and the candidate tool list shown at inference time. These interfaces often change after deployment even when the underlying tool semantics remain fixed, creating a practical reliability problem for function-calling systems. We study this failure mode as interface drift and evaluate whether a lightweight, training-free repair layer can recover robustness without retraining the base model. Our method, SchemaShield-Lite, combines schema validation, candidate-set retry for missing tool names, and one-shot constrained repair. On Qwen3.5-9B, it raises DICE-300 accuracy from 0.717 to 0.813 and BFCL-200 accuracy from 0.580 to 0.805, recovering 19/22 and 33/45 drift-induced failures with no observed repair harms. Naive retry does not explain the gains. Ablations show that candidate-set drift, especially distractor tools, accounts for most observed damage, while description and schema edits alone are less harmful. The results position validation-and-repair as a practical robustness layer for deployed LLM-agent tool use.
Modern deep learning systems increasingly rely on high-dimensional vector operations, in which large input and parameter vectors are repeatedly stored, transferred, and processed. Although substantial research has focused on reducing neural-network model size through pruning, quantization, knowledge distillation, and efficient architectures, the storage, memory-traffic, and communication costs associated with repeatedly presented high-dimensional input vectors remain important practical considerations, particularly in distributed and edge-computing environments. This paper proposes the SETAREH (Sorting-assisted Energy-concentrating Transform for Approximate Representation using Haar wavelets) transform, a Haar-wavelet-based sparsification approach for approximating the inner product between a fixed high-dimensional coefficient vector and successive input vectors. By applying the same sequence of permutations and Haar transforms to both vectors, SETAREH preserves the exact inner product before truncation. It then retains only a subset of the transformed coefficient-input pairs, reducing their representation cost while introducing a controllable approximation error. To evaluate SETAREH, we consider two complementary experimental settings: a synthetic Gaussian-vector baseline and a real apparent-age estimation task using a convolutional neural network on the APPA-REAL dataset. The results show that SETAREH can approximate the evaluated inner products using substantially fewer transformed coefficient-input pairs than the original vector dimension. The findings support its use as a potential structured data-representation method for repeated inner products with fixed weights, while the practical benefit depends on approximation accuracy and the cost of storing or transmitting permutation metadata.
Pretrained models are often reused across downstream tasks, motivating the question of whether this transfer learning procedure can provide a substantial performance improvement, namely, in terms of the scaling law. We address this question in the prototypical setting of high-dimensional regression, where (i) a pretrained source model is trained via ridge regression, and (ii) a fine-tuned target model is regularized via the L_2 distance from the source estimator. First, we establish a deterministic equivalent for the risk of this two-step procedure under both model and covariance shift. Next, by specializing the deterministic equivalent to a shared regression function and commuting power-law covariance spectra, we identify a regime where the resulting scaling law achieves the minimax-optimal exponent. Notably, in this regime, the optimal source regularization depends on source data only, indicating its universality across all targets. Furthermore, we show that the same optimal rate is achieved by performing the fine-tuning step via one-pass stochastic gradient descent. We support our theoretical analysis via experiments on both synthetic settings and language models.
The High Resolution Imaging Science Experiment (HiRISE) onboard the Mars Reconnaissance Orbiter (MRO) acquires imagery of Mars at 25--30 cm/pixel, supporting studies of martian geology, climate history, and landing-site characterization. Its 14-detector focal plane assembly includes several overlapping detectors that enable the acquisition of three-channel colour-infrared images, which are important for identifying compositional variations and studying active surface processes. In 2023, a detector in the center of the focal plane assembly (RED4) began failing intermittently, creating a gap in acquired imagery and reducing the instrument's colour-imaging capability by half. The HiRISE processing pipeline provides a synthetic red (SYN4) channel obtained by linearly scaling the co-located infrared (IR10) channel. However, this approach cannot capture observed non-linear relationships between the colour channels, limiting its accuracy and applicability. Here, we present HiPredict, a machine-learning framework that reconstructs RED4 from co-located IR10 and blue-green (BG12) imagery, neighbouring red-channel observations, and acquisition metadata. The model employs a U-Net architecture to learn non-linear spatiotemporal-spectral relationships between these inputs and is trained on 81,717 training samples, representing 13,625 observations. Evaluated on 23,348 test samples, representing 3,892 pre-failure observations, HiPredict reconstructs RED4 with a Normalised Mean Absolute Error of 1.1%, substantially outperforming SYN4 (3.1%). The reconstructed products improve colour image recovery and more closely represent features of interest such as icy craters, stratigraphy exposed along ridge crests and chloride deposits. We release HiPredict as an open software tool. This work demonstrates the potential of machine learning to mitigate data loss as planetary missions age and naturally degrade.
Real-time inference on small devices needs low activation memory without extra latency, which makes operator order the most valuable memory lever: it is free in time. We show that the usual way of deploying a better order, handing it back to the compiler and regenerating code, is unreliable. In PyTorch Inductor the regenerated program misses its predicted peak in 22 of 54 CPU configurations. It is worse than the stock program in 9 of them and in 19 of 50 on a GPU, and on the GPU it also changes numerical results. We instead profile the generated program, model its memory exactly from its own statements and their read/write effects, and permute those statements. We prove that this realizes the modeled peak whenever allocation does not depend on history. It does so to the byte, with bitwise-identical results, in all our CPU runs. On the GPU we characterize why the caching allocator breaks the condition. To know when to stop, we give forward and reverse "trap" lower bounds, each computable with O(n^2) minimum cuts. They certify the found order optimal for 94% of held-out wrapper graphs and all 16 microcontroller-class models we study, in under a second. So most inference graphs are already scheduled optimally. Branching networks are the exception, where reordering is free and pays, e.g. 10.0% on DeepLabV3. In compiled training we save up to 7.6%. Beyond the certified optimum we study selective recomputation with per-operator caps. It is strongly NP-hard for every fixed cap even at treewidth two, encodes knapsack exactly on series-parallel graphs, and is solvable exactly on bounded treewidth. It is useless on public tiny models whose peak is a single layer, but on architectures that hold a cheap value for a late reuse it pays: on the public ReDimNet speaker networks, whose order is already optimal, it lowers the peak by 5.2-6.1% at under 0.4% extra MACs, with bitwise-identical outputs.
Often, eg in healthcare and predictive maintenance, what matters is which of several competing causes will strike first, and with what probability: survival analysis with competing risks. Corresponding deep learning methods are developed as bundles: a new architecture together with a new training loss, tested on one to five datasets against a handful of baselines. Thus, a state-of-the-art result cannot be attributed to the architecture, the loss, or the evaluation protocol. The coupling is not accidental: with one incidence curve per cause, a network's head is tied to the losses expressible on it. We vary the two axes independently, three losses, two of them strictly proper, on four backbones, across two modalities that tease the axes apart: 20 tabular and 3 pathology-image datasets. One loss wins in both: modelling cumulative incidence directly under censoring weighting. It brings twice as much as benefits a good choice of backbone on competing-risks data, and no more than it on single-event data: the loss is decisive for competing risks. The backbone matters too, and must match the modality: tabular networks lead on tables, pretrained vision encoders on images. The two choices compound rather than substitute. Proper scoring rules reveal these dynamics; the concordance index (C-index), which is not proper, does not separate models: neither axis moves it, not even tuning on the loss built to improve it. This is an avenue of progress: composed, the two give a model nobody has tried, RealMLP or TabM trained with the cumulative-incidence loss, which takes the best Brier score in our benchmark.
Enterprise data analysis requires agents to connect analytical requests with knowledge distributed across databases, business documents, and existing example tasks. Without an explicit representation of this knowledge, an agent may produce incorrect analytical results and cannot reliably incorporate feedback. We present AgentOnto, an ontology-guided LLM agent for enterprise data analytics. AgentOnto organizes heterogeneous enterprise knowledge into an ontology that directly guides analysis and incorporates confirmed feedback. Iterative Ontology Preparation further improves the ontology through self-generated analysis tasks. We evaluate AgentOnto on two public data sources. AgentOnto substantially outperforms representative methods, while module-level evaluations confirm the effectiveness of ontology construction, analysis, and updating. A real-world deployment further demonstrates its practical value in enterprise analytics. These results demonstrate that explicitly constructing, using, and updating enterprise knowledge supports reliable and adaptable data analysis.
The joint reconstruction of the initial pressure and spatially varying sound speed in photoacoustic tomography from boundary measurements is considered. A regularized data-driven framework is proposed in which the unknown fields are represented in finite-dimensional wavelet spaces and fixed nonlinear features generated by an untrained convolutional neural network (CNN) are used to reconstruct the initial-pressure coefficients through Tikhonov-regularized least squares. A learned coefficient-space prior is subsequently used to infer the squared sound-speed coefficients. Numerical experiments demonstrate accurate reconstruction on held-out data and strong generalization to previously unseen phantom geometries. The frozen reconstruction map also remains stable under measurement noise not represented during training, with only minor changes in reconstruction accuracy for noise levels up to 10\%. In comparison with the supervised trained-CNN approach considered in this work, the untrained representation yields substantially smaller reconstruction errors while requiring approximately one-sixth of the computational time. These results demonstrate that fixed random convolutional features combined with regularized coefficient-space learning provide an efficient and robust approach to joint reconstruction in photoacoustic tomography.
Long-term agent memory requires recovering answer-relevant evidence dispersed across turns and sessions. Existing systems increasingly use structured memories and adaptive retrieval, yet often pass retrieved or expanded units directly to the reader, coupling memory addressing with evidence presentation and leaving the final context redundant or incomplete. To address this gap, we propose MemCompile, a query-conditioned evidence compilation framework for long-term conversational memory. Given a query, the Address Plane locates relevant support through compact source-linked descriptors and capacity-aware Cells, while a normalization boundary resolves heterogeneous retrieval hits to stable SourcePacket identities. Its Evidence Plane then allocates a bounded evidence budget across complementary packets, admits limited additional support, and renders the retained claims associated with each selected packet into an answer-ready context. MemCompile achieves 80.52% Judge Accuracy on LoCoMo and extends the empirical quality-cost Pareto frontier to a higher-quality regime. Its strong performance on LongMemEval-S further suggests that the benefits of query-conditioned compilation generalize across long-term conversational-memory benchmarks.
