Scales attention for long contexts in LLMs.
problem Development of attention mechanisms for long context inference.
method Scale-invariant total attention and sparsity conditions, with a position-dependent transformation of logits.
result Scale-invariant attention scheme improves validation loss and long-context retrieval.
The paper tackles long-context linear system identification with improved sample complexity bounds.
problem Identifying dynamical systems with long dependencies over fixed context windows.
method Established sample complexity bounds for systems with linear dependencies over a context window of length p.
result The learning process is not hindered by slow mixing properties in extended context windows.
Sparse attention model reduces long-context inference time with exponential accuracy guarantees.
problem Efficiently processing long-context queries in large language models.
method Formalizes attention as a projection onto key vectors, analyzes entropic relaxation, and introduces Vashista Sparse Attention.
result Sparse attention concentrates on a constant-size active face, leading to exponential decay of inactive tokens' mass and linear scaling of active face error.
Hydra boosts efficiency for long-context reasoning in resource-constrained settings.
problem Quadratic complexity of transformers limits long-context reasoning in resource-constrained systems.
method Hydra uses a modular architecture with adaptive routing between sparse global attention, mixture-of-experts, and dual memories.
result Hydra achieves significant throughput and accuracy improvements for long-context reasoning.
Study finds phase transition in context-sensitive language model with short-range interactions.
problem Understanding phase transitions in language models with short-range interactions.
method Constructed a random language model with short-range interactions and investigated its statistical properties.
result Phase transition occurs in context-sensitive language models with constant context length.
Transformer models waste resources on long-context tasks.
problem Redundant attention computations in Transformer models for long-context tasks.
method Reformulate sequence modeling as supervised learning, analyze attention sparsity, formulate attention optimization as linear coding problem, propose Dynamic Group Attention.
result DGA reduces computational costs while maintaining performance.
GACELA fills long gaps in musical audio with a GAN and context conditioning.
problem Restoring long gaps in musical audio with varying complexity and duration.
method Generative adversarial network (GAN) with five parallel discriminators and context conditioning.
result Reduced artifacts in inpaintings from unacceptable to mildly disturbing.
Persistent homology enhances graph classification by capturing long-range graph properties.
problem Lack of formal assessment of persistent homology in graph learning.
method Brief introduction and theoretical discussion of persistent homology in graph context, followed by empirical analysis.
result Persistent homology improves graph classification, especially for data with prominent topological structures.
A long user history inevitably reflects the transitions of personal interests over time. The analyses on the user history require the robust sequential model to anticipate the transitions and the decays of user interests. The user history is often modeled by various RNN structures, but the RNN structures in the recomme…
Timer-XL predicts multidimensional time series using a unified Transformer approach.
problem Unified time series forecasting across various tasks and contexts.
method Decoder-only Transformers with a universal TimeAttention mechanism and deft position embedding.
result State-of-the-art performance across multiple forecasting benchmarks.
In dialogues, an utterance is a chain of consecutive sentences produced by one speaker which ranges from a short sentence to a thousand-word post. When studying dialogues at the utterance level, it is not uncommon that an utterance would serve multiple functions. For instance, "Thank you. It works great." expresses bot…
Unified framework for fair pricing in long-term insurance products.
problem Unclear generalization of fair pricing methods to long-term products.
method Reformulate multi-state transition models as Poisson regression problems.
result Direct application of existing fair pricing methods to long-term insurance products.
In the sentence classification task, context formed from sentences adjacent to the sentence being classified can provide important information for classification. This context is, however, often ignored. Where methods do make use of context, only small amounts are considered, making it difficult to scale. We present a …
Stable Hadamard Memory improves reinforcement learning by efficiently managing memory.
problem Memory models struggle in partially observable reinforcement learning environments.
method Introduces a novel memory model using the Hadamard product for efficient memory management and updates.
result Significantly outperforms state-of-the-art memory-based methods on challenging benchmarks.
Mamba struggles with long context lengths, but spectrum scaling improves performance.
problem Mamba's performance degrades with increasing context length.
method Spectrum scaling applied to pre-trained Mamba models to improve long-context generalization.
result Spectrum scaling significantly improves performance in long-context settings.
MPPN network improves long-term time series forecasting accuracy.
problem Inaccurate long-term time series forecasting due to noise and lack of interpretability.
method MPPN network constructs context-aware multi-resolution semantic units and employs multi-periodic pattern mining and channel adaptive module.
result MPPN significantly outperforms state-of-the-art methods on nine real-world benchmarks.
New Performer model tackles long-sequence protein modeling.
problem Challenges of training complex Transformer models for long sequences.
method Linearly scalable long-context Transformer architecture, Performer.
result Performer provides strong theoretical guarantees and is effective for protein sequence modeling.
Understanding temporal dynamics has proved to be highly valuable for accurate recommendation. Sequential recommenders have been successful in modeling the dynamics of users and items over time. However, while different model architectures excel at capturing various temporal ranges or dynamics, distinct application cont…
Transformers recall from long distributions with statistical guarantees.
problem Designing Transformers that can recall from arbitrarily long, distributional contexts.
method Recast associative memory as probability measures, decomposing the task into recall and prediction.
result A shallow measure-theoretic Transformer learns the recall-and-predict map under spectral assumptions.
We consider the task of detecting regulatory elements in the human genome directly from raw DNA. Past work has focused on small snippets of DNA, making it difficult to model long-distance dependencies that arise from DNA's 3-dimensional conformation. In order to study long-distance dependencies, we develop and release …
Short-term demand forecasting models commonly combine convolutional and recurrent layers to extract complex spatiotemporal patterns in data. Long-term histories are also used to consider periodicity and seasonality patterns as time series data. In this study, we propose an efficient architecture, Temporal-Guided Networ…
Context-aware recommender systems (CARSs) apply sensing and analysis of user context in order to provide personalized services. Adding context to a recommendation model is challenging, since the addition of context may increases both the dimensionality and sparsity of the model. Recent research has shown that modeling …
Study shows short exposure and systematic risk exposure affect disposition effect asymmetries.
problem Understanding disposition effect in short vs long exposure positions and systematic risk.
method Generalized Odean measures, introduced Value metric, implemented dispositionEffect R package.
result Short positions exhibit weaker disposition effect than long positions under narrow framing, reversing in integrated framing.
One of the primary challenges of visual storytelling is developing techniques that can maintain the context of the story over long event sequences to generate human-like stories. In this paper, we propose a hierarchical deep learning architecture based on encoder-decoder networks to address this problem. To better help…
Rough Transformers improve efficiency for medical time-series data.
problem Efficiently modeling irregularly sampled, long-range time-series data.
method Introducing Rough Transformers, a Transformer variant with continuous-time representations and multi-view signature attention.
result Rough Transformers outperform vanilla Transformers while using less computational resources.
In a voice-controlled smart-home, a controller must respond not only to user's requests but also according to the interaction context. This paper describes Arcades, a system which uses deep reinforcement learning to extract context from a graphical representation of home automation system and to update continuously its…
In this paper we examine a possible reason for the LSTM outperforming the GRU on language modeling and more specifically machine translation. We hypothesize that this has to do with counting. This is a consistent theme across the literature of long term dependence, counting, and language modeling for RNNs. Using the si…
A new algorithm reduces memory usage for long token attention in streaming applications.
problem Memory inefficiency in computing attention for long documents.
method One-pass streaming algorithm using sublinear space storage.
result Super-efficient memory usage for long token attention.
The paper proposes Tier Balancing for dynamic fairness in decision-making.
problem Achieving long-term fairness in decision-making processes.
method Causal modeling with DAGs to investigate dynamic fairness.
result Tier Balancing is a more natural approach to achieve long-term fairness, capturing latent causal factors.
Statistical test verifies long-term rating system calibration with overlapping time windows.
problem Verifying supervisory requirements for overlapping time windows in rating systems.
method Analyzes long-run default rate distribution and correlation effects; presents conservative calibration test methods.
result Developed a test for individual and portfolio levels that can handle unknown variance.
FibQuant improves KV-cache compression for long-context inference.
problem Memory traffic bottleneck in long-context inference due to KV cache growth.
method Introduces FibQuant, a universal vector quantizer that combines Beta-quantile radii, Fibonacci/Roberts-Kronecker directions, and Lloyd-Max refinement.
result FibQuant achieves high compression rates with minimal loss in attention cosine similarity.
The study examines how extra compute during testing affects the performance of large language models.
problem Understanding the conditions under which test-time scaling improves model performance.
method An in-context weight prediction task for linear regression was used to train transformers. The performance was analyzed under varying levels of test-time compute.
result Training transformers on diverse, relevant, and hard tasks leads to the best performance for test-time scaling.
We describe a new class of learning models called memory networks. Memory networks reason with inference components combined with a long-term memory component; they learn how to use these jointly. The long-term memory can be read and written to, with the goal of using it for prediction. We investigate these models in t…
We employ perturbation analysis technique to study multi-asset portfolio optimisation with transaction cost. We allow for correlations in risky assets and obtain optimal trading methods for general utility functions. Our analytical results are supported by numerical simulations in the context of the Long Term Growth Mo…
Models for audio source separation usually operate on the magnitude spectrum, which ignores phase information and makes separation performance dependant on hyper-parameters for the spectral front-end. Therefore, we investigate end-to-end source separation in the time-domain, which allows modelling phase information and…
Existing language models such as n-grams for software code often fail to capture a long context where dependent code elements scatter far apart. In this paper, we propose a novel approach to build a language model for software code to address this particular issue. Our language model, partly inspired by human memory, i…
The paper addresses estimating long-term treatment effects with monotone missing data.
problem Estimating long-term treatment effects with missing data, especially monotone missing.
method The paper introduces the sequential missingness assumption for identification and proposes three novel estimation methods: inverse probability weighting, sequential regression imputation, and SeqMSM. It also introduces a balancing-enhanced approach, BalanceNet, to improve estimation accuracy.
result The proposed methods, including BalanceNet, effectively estimate long-term treatment effects with monotone missing data.
Express improves causal attention guarantees for language models.
problem Improving causal attention guarantees for language models.
method Introducing Express, a tool for converting non-causal attention into causal with matching guarantees.
result Express improves causal attention guarantees to log^(3/2)(n)/s with minimal memory and compression overhead.
Transformers learn rich in-context dependencies efficiently.
problem Understanding how transformers learn long-range dependencies efficiently.
method Approximation and dynamics analysis of induction head mechanisms.
result Abrupt transition from lazy to rich mechanisms during training.
Bayesian context trees capture complex dependencies in categorical sequences.
problem Complex, long-range dependencies in categorical sequences are not well captured by simple models.
method Parsimonious Bayesian context trees with model-based agglomerative clustering for efficient inference.
result The proposed framework outperforms existing models on real-world data.
Proposes a new RNN for language generation capturing long-range dependencies.
problem Capturing long-range word dependencies and sentence order in text corpora.
method Recurrent Hierarchical Topic-Guided RNN with dynamic deep topic model.
result Outperforms larger-context RNN-based language models and learns interpretable topics.
Combines CNN and Transformer for financial time series forecasting.
problem Forecasting financial time series, especially stock prices, is challenging due to short-term and long-term dependencies.
method Uses CNN for short-term dependencies and Transformer for long-term dependencies.
result Demonstrated superior performance in forecasting stock price changes compared to traditional methods.
Attention-only transformers learn from context via two stages of inference.
problem Learning from corrupted token sequences in minimal transformers.
method Two-stage empirical Bayes interpretation: kernel-weighted posterior mean and particle dynamics.
result Effective denoising without explicit noise schedules, showing posterior-mean recovery under asymptotic conditions.
Estimates long-term effects from short-term experiments and observational data with unobserved confounders.
problem Estimating long-term causal effects from short-term experiments and long-term observational data with unobserved confounding.
method Combining regression residuals with short-term experimental outcomes to create an instrumental variable for estimating long-term causal effects.
result The estimator is unbiased and its variance is analytically studied.
Self-attention prefers sparse functions of input sequences, reducing sample complexity.
problem Understanding the inductive biases of self-attention in modeling long-range dependencies.
method Theoretical analysis and synthetic experiments to probe sample complexity of learning sparse functions with Transformers.
result Bounded-norm Transformer networks can represent sparse functions of the input sequence with logarithmic sample complexity.
RNNs struggle with in-context retrieval, while Transformers excel.
problem In-context retrieval capability of RNNs.
method Theoretical analysis and experimental techniques (CoT, RAG, Transformer layer).
result Enhancing RNNs with techniques improves their in-context retrieval capability, closing the representation gap with Transformers.
Paper addresses hidden faces in configuration space integrals for embeddings.
problem Understanding hidden faces in configuration space integrals for long embeddings.
method Modified configuration space integrals incorporating acyclic bar complex of a dg algebra.
result Cochain map from new graph complex to de Rham complex of embeddings modulo immersions.
Most real life systems have a random component: the multitude of endogenous and exogenous factors influencing them result in stochastic fluctuations of the parameters determining their dynamics. These empirical systems are in many cases subject to noise of multiplicative nature. The special properties of multiplicative…