New methods improve integration of external LMs with AED models.
problem Improving performance of AED models by integrating external LMs.
method Comparing and proposing novel methods to estimate implicit LM from AED models.
result Proposed methods outperform previous approaches.
Language models perform worse with implicit reward models than explicit ones.
problem Understanding why implicit reward models generalize worse than explicit ones.
method Investigated the root cause of the generalization gap between IM-RMs and EX-RMs.
result Implicit reward models rely more on superficial token-level cues, leading to worse generalization.
Models learn spatial templates from implicit language, predicting spatial arrangements with high accuracy.
problem Predicting spatial arrangements from implicit spatial language.
method Simple neural-based models leveraging annotated images and structured text.
result Models can predict spatial arrangements from implicit spatial language with high accuracy, even for unseen objects.
Research aims to ensure fair classification across explicit and implicit sensitive features.
problem Ensuring fairness in machine learning models when sensitive features are not explicitly provided.
method Defined explicit and implicit cohorts, used clustering of embeddings, modified loss function.
result Improved classification parity across explicit and implicit sensitive features.
The study provides a generalization bound for a family of implicit networks.
problem Theoretical understanding of implicit networks' generalization is limited.
method A generalization bound is derived for a family of implicit networks using a covering number argument for Rademacher complexity.
result A theoretical generalization bound is established for implicit networks.
Improved speech recognition with language model integration in sequence-to-sequence models.
problem Improving word error rate in speech recognition models.
method Log-linear combination of acoustic and language models with per-token renormalization.
result The proposed method shows good improvements over standard model combination on Librispeech system.
Gradient descent in deep matrix factorization favors low-rank solutions, improving recovery accuracy.
problem Understanding the generalization in deep learning models.
method Study of gradient descent over deep linear neural networks for matrix completion and sensing.
result Adding depth enhances an implicit tendency towards low-rank solutions, leading to more accurate recovery.
Improved MUSE boosts performance and reduces error in Bayesian inference.
problem Hierarchical Bayesian inference problems
method Implicit differentiation applied to MUSE algorithm
result Significant speedup and improved accuracy compared to Hamiltonian Monte Carlo
New method shows cultural transmission improves language compositionality.
problem How cultural transmission affects language emergence and compositionality.
method Introduced cultural evolutionary dynamics into language emergence by periodically replacing agents in a population.
result Languages emerge with better compositional generalization due to implicit cultural transmission.
Amobee wins WASSA 2018 emotion prediction with language models and LSTM.
problem Predicting emotions from tweets without explicit mentions.
method Ensemble system of language models and LSTM with CNN attention.
result 1st place with macro F1 score of 0.7145.
SAM improves deep learning tasks by promoting balancedness, reducing outlier impact.
problem Improving generalization in deep learning tasks, especially with scale-invariant problems.
method Introduces balancedness as a new concept to depict global behaviors of SAM, focusing on the difference between squared norms of two variables.
result SAM promotes balancedness and is data-responsive, outperforming SGD in outlier scenarios.
RevDEQs improve performance on tasks with exact gradients and fewer function evaluations.
problem Inexact gradient calculation in DEQs leads to unstable training and requires regularisation or many function evaluations.
method Introduce Reversible Deep Equilibrium Models (RevDEQs) that allow for exact gradient calculation, no regularisation, and far fewer function evaluations.
result RevDEQs significantly improve performance on language modelling and image classification tasks.
Adaptive scheduling improves multilingual neural machine translation models.
problem Training models on multiple tasks with uniform or proportional sampling leads to poor performance trade-offs.
method Exploring non-adaptive and adaptive task scheduling, including implicit schedules.
result Adaptive schedules improve model performance for low-resource tasks without negatively affecting high-resource tasks.
Proposes EM-HRNN model for better hierarchical language representation.
problem Learning hierarchical representations in natural language processing.
method Introduces latent indicator layer and EM algorithm for training.
result EM-HRNN model outperforms other RNN-based models in document classification.
EXPO framework eliminates need for reward model, achieving better optimization.
problem Optimizing large language model responses without a separate reward model.
method Introduces EXPO framework that avoids reparameterization, directly optimizing preferences.
result Demonstrates better regularization and intuitive interpolation behaviors.
This work explains how linear representations in large language models arise from training objectives and gradient descent.
problem Understanding the origins of linear representations in large language models.
method A latent variable model to abstract and formalize concept dynamics, combined with analysis of the softmax cross-entropy objective and gradient descent.
result Linear representations emerge when learning from data matching the latent variable model, and this simple structure suffices to yield linear representations.
Advances deep latent variable models for more flexible text generation.
problem Limited representation power of VAEs due to Gaussian assumptions and posterior collapse.
method Develops sample-based variational distributions and an LVM to directly match aggregated posterior to prior.
result Demonstrates improved text generation in various scenarios.
Gradient flow with infinitesimal initialization converges to Greedy Low-Rank Learning for matrix factorization.
problem Understanding implicit regularization in gradient descent for matrix factorization.
method Theoretical and empirical analysis of gradient flow with infinitesimal initialization and Greedy Low-Rank Learning.
result Gradient flow with infinitesimal initialization is mathematically equivalent to Greedy Low-Rank Learning for depth-2 matrix factorization under reasonable assumptions.
SGD noise has no bias advantage in online learning, contrary to offline learning.
problem The role of SGD noise in online learning.
method Extensive empirical analysis of image and language data.
result Small batch sizes do not confer any implicit bias advantages in online learning.
MMVAE learns multi-modal data with shared and private latent spaces.
problem Learning useful representations across multiple data modalities.
method Mixture-of-experts variational autoencoder (MMVAE).
result MMVAE satisfies four criteria for multi-modal learning.
Model predicts increased social unrest during COVID-19 using social media data.
problem Detecting rising conflict potential in societies during pandemics.
method Neural implicit motive pattern recognition from social media texts.
result Significant increase in conflict indicators during the pandemic.
This work analyzes how transformers learn common linear regression tasks.
problem Understanding how in-context learning operates in real-world applications with common task structures.
method Analyzing a linear attention model trained on low-rank regression tasks.
result Statistical fluctuations in finite pre-training data induce an implicit regularization, leading to a sharp phase transition in generalization error.
This paper presents a novel latent variable recurrent neural network architecture for jointly modeling sequences of words and (possibly latent) discourse relations between adjacent sentences. A recurrent neural network generates individual words, thus reaping the benefits of discriminatively-trained vector representati…
Graph Structured Prediction Energy Networks model correlations for joint inference.
problem Joint inference over multiple variables with high-order correlations.
method Energy Networks for modeling explicit local and implicit higher-order correlations.
result Tractable inference with explicit modeling of correlations.
Neuro-symbolic system tackles conversational AI's need for natural, broad-ranging dialogue.
problem Understanding unstated presumptions in conversational AI commands.
method Neuro-symbolic theorem prover for multi-hop reasoning.
result Extracts multi-hop reasoning chains from natural language commands.
This work fine-tunes GPT-2 for generating patent claims.
problem Generating coherent patent claims automatically.
method Fine-tuning OpenAI GPT-2 on patent claim language structure.
result Demonstrated the first machine-generated patent claims.
This study analyzes how one-layer transformers learn regular language recognition tasks.
problem Understanding how one-layer transformers solve regular language recognition tasks like even pairs and parity check.
method Theoretical analysis of training dynamics and gradient descent for a one-layer transformer.
result A one-layer transformer can solve even pairs directly but needs CoT for parity check. Training phases show rapid growth in attention layer followed by logarithmic growth in linear layer.
New principles needed for scaling large language models, challenging traditional regularization methods.
problem The shift from generalization to scaling in machine learning requires new guiding principles.
method Examining the effectiveness of traditional regularization methods in the scaling-centric era.
result Traditional principles of regularization may not generalize to larger scales, highlighting new phenomena like scaling law crossover.
Paper proposes f-DPG for aligning language models with preferences.
problem Aligning language models with user preferences.
method Uses f-divergence to approximate target distributions and minimizes a forward KL from it using DPG.
result Jensen-Shannon divergence often outperforms forward KL divergence, leading to significant improvements.
LLMs encode latent topic distributions, suggesting Bayesian inference.
problem Capturing topic structure from large language models.
method Connecting LLM optimization to implicit Bayesian inference and de Finetti's theorem.
result LLMs recover latent topic distributions, matching LDA-generated topics.
New approach predicts tokens in context, explaining how ICL emerges.
problem Limited understanding of in-context learning emergence.
method Auto-regressive next-token prediction (AR-NTP) with prompt token-dependency and a two-level expectation.
result ICL emerges from the generalization of sequences and topics.
Zeroth-order methods favor flat minima in machine learning.
problem Finding solutions with small Hessian trace in optimization.
method Zeroth-order optimization with two-point estimator.
result Zeroth-order optimization converges to flat minima.
Efficiently distills white-box adversarial attacks into black-box models.
problem Generating efficient adversarial examples for robustness.
method Train a model to emulate white-box attack behavior and distill it into a more efficient black-box model.
result Reduces adversarial example generation time by 19x-39x and transfers to black-box settings.
XPO enhances RLHF by encouraging diverse responses, offering improved sample efficiency.
problem Limited exploration in RLHF leads to suboptimal models and bottlenecks.
method XPO is a simple one-line change to DPO, introducing a novel exploration bonus.
result XPO achieves strong theoretical guarantees and promising empirical performance.
Tensor models decode human perception and memory using SPO triples.
problem Understanding implicit and explicit perception and memory in the brain.
method Tensor models with SPO triples, dual representations, and four layers.
result Semantic memory is crucial for explicit perception and declarative memories.
This paper shows how to train only the implicit layer of overparameterized implicit neural networks.
problem Understanding how the implicit layer contributes to the training of overparameterized implicit neural networks.
method Restricting training to only the implicit layer and analyzing the generalization error for ReLU-activated networks.
result Global convergence is guaranteed even if only the implicit layer is trained, and gradient flow with proper random initialization can achieve small generalization errors.
Aioli unifies language model data mixing methods and improves performance.
problem Optimizing the mixture of training data groups for language models.
method Unified optimization framework for dynamically adjusting mixture proportions.
result Aioli outperforms existing methods by up to 12.012 test perplexity points.
AdamW optimizes a constrained loss with ℓ∞ norm constraint.
problem Understanding the optimization behavior of AdamW with ℓ∞ norm constraint. method Analyzing AdamW as a smoothed version of SignGD and connecting it to Frank-Wolfe optimization.
result AdamW implicitly performs constrained optimization with ℓ∞ norm constraint. DEQ models find equilibrium points for efficient deep learning.
problem Efficiently modeling sequential data with reduced memory usage.
method Directly find equilibrium points in an infinite depth feedforward network using implicit differentiation.
result DEQs often improve performance and reduce memory consumption compared to existing models.
Paper studies the theoretical equivalence between implicit and explicit neural networks in high dimensions.
problem Lack of theoretical analysis of implicit and explicit neural networks.
method Examined high-dimensional implicit neural networks and established their equivalence to explicit networks.
result Equivalence between implicit and explicit neural networks in high dimensions.
Unified approach to RLHF tackles uncertainty in reward function.
problem Uncertainty in reward function learned from human feedback.
method Value-incentivized preference optimization (VPO) that regularizes the reward function with value function.
result Theoretical and practical guarantees for both online and offline RLHF settings.
Gradient matching method estimates implicit regularization in complex deep learning systems.
problem Estimating implicit regularization in modern deep learning systems with complex modifications.
method Gradient matching methods to empirically estimate implicit regularization.
result Empirical estimation of implicit regularization in arbitrary networks, including dropout.
VAEs improve collaborative filtering for implicit feedback.
problem Limited modeling capacity of linear factor models in collaborative filtering.
method Introduced a generative model with multinomial likelihood and used Bayesian inference for parameter estimation.
result Significantly outperforms state-of-the-art baselines on real-world datasets.
LLMs generate answers under incomplete context, and their uncertainty should scale with missing information.
problem Evaluating the quality of LLM answers under incomplete context.
method A controlled framework with varying context availability, and two uncertainty measures (sampling-based confidence and response entropy) evaluated on SQuAD.
result Response entropy increases with context removal and explains more variance in accuracy than confidence, suggesting it is a more responsive uncertainty measure.
Continuous semi-implicit models enable faster training and better performance in generative modeling.
problem Slow convergence in hierarchical semi-implicit models during training.
method CoSIM, a continuous semi-implicit model that incorporates a continuous transition kernel for efficient training.
result CoSIM achieves superior performance on image generation tasks compared to existing methods.
Gradient descent converges to a global minimum in nonlinear ReLU implicit networks with linear width.
problem Understanding convergence of gradient methods in nonlinear, infinitely deep ReLU networks.
method Introduced a scaling constant to ensure well-posedness of the equilibrium equation, proving convergence to a global minimum for linear width networks.
result Gradient descent converges to a global minimum at a linear rate for nonlinear ReLU implicit networks with linear width.
Gradient descent biases linear models in next-token prediction towards data entropy.
problem Optimization bias in next-token prediction models.
method Analysis of gradient descent on linear models with sparse conditional distributions.
result Gradient descent selects parameters that equate token logits differences to log-odds in the data subspace.
DSIVI improves variational autoencoders by optimizing a proper lower bound on ELBO.
problem Improving variational autoencoders with implicit priors.
method Introducing DSIVI, a method that optimizes a proper lower bound on ELBO for models with semi-implicit priors and posteriors.
result DSIVI improves the performance of VampPrior, a state-of-the-art prior for variational autoencoders.