Proposes a simple framework to balance task difficulty in multi-task learning.
problem Varying difficulty levels among different tasks in multi-task learning.
method Introduces a Balanced Multi-Task Learning (BMTL) framework that transforms training losses to balance task difficulty.
result Empirical studies show state-of-the-art performance of the proposed BMTL framework.
New measure quantifies task difficulty for machine learning models.
problem Quantifying the inherent difficulty of machine learning tasks.
method Inductive bias complexity measure.
result Tasks requiring generalization over many dimensions are more difficult.
Study shows perceptual boost of visual attention varies with task difficulty and size.
problem Understanding how task-dependent the perceptual boost of visual attention is in natural settings.
method Designed and trained neural networks on various visual tasks, comparing results to baseline.
result Perceptual boost of attention is stronger with more difficult tasks and weaker with larger task sets.
Proposes PIC and POIC for measuring task difficulty in RL.
problem Lack of metrics to measure task difficulty in RL.
method Introduces policy information capacity (PIC) and policy-optimal information capacity (POIC) as metrics based on mutual information.
result Empirically shows PIC and POIC correlate with task solvability better than alternatives.
Entity Linking (EL) is the task of automatically identifying entity mentions in a piece of text and resolving them to a corresponding entity in a reference knowledge base like Wikipedia. There is a large number of EL tools available for different types of documents and domains, yet EL remains a challenging task where t…
This study introduces balanced DRPS and OrderedLogitNN for better QDE of discrete-level questions.
problem Lack of ordinal regression methods and fair evaluation metrics for discrete-level QDE.
method Introduces balanced DRPS and OrderedLogitNN, fine-tunes BERT on RACE++ and ARC datasets.
result OrderedLogitNN outperforms other models on complex QDE tasks.
Transformers learn to adapt to different task difficulties and resist distribution shifts.
problem Understanding and optimizing a Transformer's performance across various task difficulties and distribution shifts.
method Analyzing a pretrained Transformer on a mixture distribution of tasks, proving optimal convergence rates.
result Transformers achieve optimal convergence rates on tasks of specific difficulty levels, robust to distribution shifts.
The paper explores when to prioritize easy or hard samples in learning tasks.
problem Determining the optimal order of learning easy or hard samples.
method Theoretical analyses and experiments were conducted to propose and validate four priority modes.
result Four priority modes (easy-first, hard-first, medium-first, two-ends-first) can be flexibly applied.
Project explores reinforcement learning solutions for sparse reward environments.
problem Difficulty in navigating environments with infrequent rewards.
method Contrast and investigate existing reinforcement learning solutions in various video games.
result Introduces a novel reinforcement learning solution combining curiosity and auxiliary tasks.
VLM judges rank well but score poorly; task difficulty and annotation quality affect interval width.
problem VLMs as judges lack reliability indicators in multimodal evaluations.
method Conformal prediction using score-token log-probabilities.
result Evaluation uncertainty is task-dependent, affecting interval width and reliability.
Algorithm learns to solve tasks by gradually expanding policy space.
problem Reinforcement learning difficulty progression.
method Decreasingly constrained policy space expansion.
result Superior learning rate in Tetris tasks.
Time-continuous emotion prediction has become an increasingly compelling task in machine learning. Considerable efforts have been made to advance the performance of these systems. Nonetheless, the main focus has been the development of more sophisticated models and the incorporation of different expressive modalities (…
CLOPS improves deep learning for continuous physiological data.
problem Deep learning struggles with streaming, multi-sensor clinical data.
method CLOPS is a replay-based continual learning strategy.
result CLOPS outperforms state-of-the-art methods in continual learning scenarios.
Reinforcement learning is a promising approach to developing hard-to-engineer adaptive solutions for complex and diverse robotic tasks. However, learning with real-world robots is often unreliable and difficult, which resulted in their low adoption in reinforcement learning research. This difficulty is worsened by the …
Paper defines a new distance metric for comparing learning tasks.
problem Comparing difficulty of learning tasks between source and target.
method Information geometry, optimal transport, coupled transfer distance.
result Coupled transfer distance correlates with fine-tuning difficulty.
In the present work we propose a Deep Feed Forward network architecture which can be trained according to a sequential learning paradigm, where tasks of increasing difficulty are learned sequentially, yet avoiding catastrophic forgetting. The proposed architecture can re-use the features learned on previous tasks in a …
New method learns optimal environment and goal difficulty for reinforcement learning.
problem Lack of robust transfer in reinforcement learning.
method Self-Supervised Active Domain Randomization (SS-ADR).
result Optimal domain randomization strategy improves transfer in goal-directed tasks.
In this work, we introduce the concept of bandlimiting into the theory of machine learning because all physical processes are bandlimited by nature, including real-world machine learning tasks. After the bandlimiting constraint is taken into account, our theoretical analysis has shown that all practical machine learnin…
Differentiable ABMs face challenges in inference and optimisation.
problem Challenges in parameter inference and optimisation for differentiable ABMs.
method Discussion and experiments highlighting challenges.
result Challenges remain in constructing differentiable ABMs.
The paper develops a theory for iterative self-improvement of models, proving conditions for better performance with easy-to-hard curricula.
problem Lack of theoretical foundation for iterative self-improvement in practical settings.
method Modeling self-improvement as maximum-likelihood fine-tuning on reward-filtered distributions and proving finite-sample guarantees.
result Explicit feedback loop and conditions for better performance with easy-to-hard curricula.
Curriculum learning strategies improve GAN training speed and quality.
problem Training GANs is notoriously difficult and time-consuming.
method Proposes three curriculum learning strategies based on difficulty scores.
result Curriculum learning strategies lead to faster convergence and superior results.
This report contains the details regarding our submission to the OffensEval 2019 (SemEval 2019 - Task 6). The competition was based on the Offensive Language Identification Dataset. We first discuss the details of the classifier implemented and the type of input data used and pre-processing performed. We then move onto…
Paper proves impossibility of three desirable properties in node embedding.
problem Understanding limitations of node embedding methods.
method Axiomatic approach to node embedding, proving impossibility of three properties.
result No node embedding method can satisfy all three desirable properties simultaneously.
A new method estimates uncertainty without explicit prediction models.
problem Costly data acquisition in machine learning.
method Distance-weighted Class Impurity method for uncertainty estimation.
result Distance-weighted Class Impurity effectively estimates uncertainty without prediction models.
Paper recovers top-two answers and confusion probability in multi-choice crowdsourcing.
problem Recovering top-two answers and confusion probability in multi-choice crowdsourcing tasks.
method Proposes a two-stage inference algorithm based on a model quantifying task difficulty and worker reliability.
result Achieves minimax optimal convergence rate and outperforms other algorithms in synthetic and real data experiments.
Learning-to-learn or meta-learning leverages data-driven inductive bias to increase the efficiency of learning on a novel task. This approach encounters difficulty when transfer is not advantageous, for instance, when tasks are considerably dissimilar or change over time. We use the connection between gradient-based me…
CogScale benchmarks AI architectures for sequential processing.
problem Evaluating AI architectures' ability to process sequential information efficiently.
method 14 scalable synthetic tasks designed to isolate cognitive and memory abilities at different scales.
result Attention mechanisms and modern state-space models consistently maintain high performance as task difficulty scales.
Paper provides conditions for reliable use of pre-trained embeddings in econometrics.
problem Uncertainty in using pre-trained embeddings for econometric tasks.
method Derives sufficient conditions and convergence rates for machine learning models with pre-trained embeddings.
result Establishes theoretical foundations for reliable use of pre-trained embeddings in econometrics.
Single-stage neural architecture improves text-to-image synthesis.
problem Combining text and image generation is challenging.
method Used deep residual networks and sentence interpolation.
result Achieved state-of-the-art performance with single-stage training.
Learning shrinks hard tail, improving inference performance.
problem Improving inference performance in neural networks.
method Latent Instance Difficulty (LID) model analyzing fine-tuning of neural networks.
result Training-dependent inference scaling, with βexteff growing with sample size before saturating. AER dynamically adjusts entropy regularization for better LLM reinforcement learning.
problem Policy entropy collapse in RLVR training limits exploration and reasoning performance.
method Adaptive Entropy Regularization (AER) with difficulty-aware coefficient allocation, initial-anchored target entropy, and dynamic global coefficient adjustment.
result AER consistently outperforms baselines on mathematical reasoning benchmarks, improving both accuracy and exploration.
We propose an active learning method for discovering low-dimensional structure in high-dimensional Gaussian process (GP) tasks. Such problems are increasingly frequent and important, but have hitherto presented severe practical difficulties. We further introduce a novel technique for approximately marginalizing GP hype…
The paper studies how to allocate human validation in AI-assisted tasks to minimize errors.
problem Heterogeneous reliability of AI-generated signals across tasks, products, and customer segments.
method Tuned prediction-powered inference, upper confidence bounds policy, Neyman square-root rule.
result The proposed policy outperforms uniform and epsilon-greedy allocation, closing most of the gap to the oracle when reliability is heterogeneous.
One of the main difficulties in analyzing neural networks is the non-convexity of the loss function which may have many bad local minima. In this paper, we study the landscape of neural networks for binary classification tasks. Under mild assumptions, we prove that after adding one special neuron with a skip connection…
GANs struggle with density estimation on simple datasets, while normalizing flows perform well.
problem Comparing GANs and normalizing flows for density estimation on non-image data.
method An extensive grid search over GAN architectures, hyperparameters, and training procedures; comparison on low-dimensional synthetic datasets.
result Normalizing flows outperform GANs in density estimation, especially in terms of Wasserstein distance.
Experience replay is an important technique for addressing sample-inefficiency in deep reinforcement learning (RL), but faces difficulty in learning from binary and sparse rewards due to disproportionately few successful experiences in the replay buffer. Hindsight experience replay (HER) was recently proposed to tackle…
Paper tackles medical image diagnosis with unsupervised domain adaptation.
problem Limited labeled samples and label noise in medical images.
method Collaborative Unsupervised Domain Adaptation (UDA) algorithm.
result Empirical results show superiority of the proposed method.
We propose a novel approach for estimating the difficulty and transferability of supervised classification tasks. Unlike previous work, our approach is solution agnostic and does not require or assume trained models. Instead, we estimate these values using an information theoretic approach: treating training labels as …
New framework optimizes deep learning training by deferring large batch sizes to late stages.
problem Optimizing batch size scheduling for deep learning training efficiency.
method Introduced the functional scaling law (FSL) framework to analyze and optimize batch size scheduling.
result Large batch sizes can be deferred to late training stages without sacrificing performance.
Policy optimization on high-dimensional continuous control tasks exhibits its difficulty caused by the large variance of the policy gradient estimators. We present the action subspace dependent gradient (ASDG) estimator which incorporates the Rao-Blackwell theorem (RB) and Control Variates (CV) into a unified framework…
Shaping in humans and animals has been shown to be a powerful tool for learning complex tasks as compared to learning in a randomized fashion. This makes the problem less complex and enables one to solve the easier sub task at hand first. Generating a curriculum for such guided learning involves subjecting the agent to…
Neural network learns its size and structure during training.
problem Adapting neural network architecture to specific datasets.
method Flexible setup allowing neural network to learn size and topology during training.
result Trained networks achieve virtually identical performance and have learned optimal structure.
PCHID improves sample efficiency in reinforcement learning tasks.
problem Sparse rewards make learning policies difficult in reinforcement learning.
method Hindsight Inverse Dynamics with Hindsight Experience Replay and Policy Continuation.
result PCHID significantly improves sample efficiency and final performance on multi-goal tasks.
This paper introduces Task 2 of the DCASE2019 Challenge, titled "Audio tagging with noisy labels and minimal supervision". This task was hosted on the Kaggle platform as "Freesound Audio Tagging 2019". The task evaluates systems for multi-label audio tagging using a large set of noisy-labeled data, and a much smaller s…
Reward shaping is one of the most effective methods to tackle the crucial yet challenging problem of credit assignment in Reinforcement Learning (RL). However, designing shaping functions usually requires much expert knowledge and hand-engineering, and the difficulties are further exacerbated given multiple similar tas…
Improves model calibration and selection in unsupervised domain adaptation.
problem Distribution shifts in unsupervised domain adaptation.
method Developed a novel importance weighted group accuracy estimator.
result Improves state-of-the-art performances by 22% in model calibration and 14% in model selection.
Deep networks prioritize easier examples over harder ones, leading to faster training.
problem Understanding how deep networks prioritize examples of varying difficulty.
method Investigated the effect of linear vs non-linear learning modes on example difficulty.
result Non-linear dynamics tend to sequentialize the learning of examples of increasing difficulty.
Dance Dance Revolution (DDR) is a popular rhythm-based video game. Players perform steps on a dance platform in synchronization with music as directed by on-screen step charts. While many step charts are available in standardized packs, players may grow tired of existing charts, or wish to dance to a song for which no …