Neural Information Processing Systems http://nips.cc/

Vision-Language Models (VLMs) suffer from catastrophic forgetting when sequentially fine-tuned on new tasks, degrading performance on previously learned foundational and task-specific capabilities. While multi-task learning can mitigate forgetting, it requires simultaneous access to all datasets and imposes computational overhead that scales linearly with the number of tasks. In this work, we introduce a routing-based approach that enables the integration of new tasks while preserving the foundational knowledge acquired during pretraining. We evaluate our method using InternVL-2 models (2B and 8B parameters) and demonstrate that routing preserves the model's foundational capabilities by maintaining performance on general-purpose benchmarks such as ChartQA, MMBench, and DocVQA, while simultaneously improving accuracy on specialized tasks. Importantly, our approach achieves this without requiring concurrent access to data from all tasks, avoiding the significant computational and data overhead associated with traditional multi-task learning. We further conduct extensive ablation studies to evaluate the scalability and robustness of routing-based learning, showing that the approach is resilient to a growing number of tasks and performs particularly well when new tasks are semantically related. Finally, we show that the routing mechanism enables superior cross-modal transfer between language and vision capabilities, allowing knowledge learned in one modality to enhance performance in another capability not achieved by existing continual learning methods.

In this paper, we present our submission for the token prediction task of EvaCun 2025. Our sys-tems are based on LLMs (Command-R, Mistral, and Aya Expanse) fine-tuned on the task data provided by the organizers. As we only pos-sess a very superficial knowledge of the subject field and the languages of the task, we simply used the training data without any task-specific adjustments, preprocessing, or filtering. We compare 3 different approaches (based on 3 different prompts) of obtaining the predictions, and we evaluate them on a held-out part of the data.

A proper understanding of the striking generalization abilities of deep neural networks presents an enduring puzzle.

Neural Information Processing Systems http://nips.cc/

Lifelong Reinforcement Learning (LRL) holds significant potential for addressing sequential tasks, but it still faces considerable challenges. A key difficulty lies in effectively preventing catastrophic forgetting and facilitating knowledge transfer while maintaining reliable decision-making performance across subsequent tasks in dynamic environments. To tackle this, we propose a novel framework, SDW (Similarity-Driven Weighting Framework), which leverages large-language-model-generated dynamic functions to precisely control the training process. The core of SDW lies in two functions pre-generated by large models: the task similarity function and the weight computation function. The task similarity function extracts multidimensional features from task descriptions to quantify the similarities and differences between tasks in terms of states, actions, and rewards. The weight computation function dynamically generates critical training parameters based on the similarity information, including the proportion of old task data stored in the Replay Buffer and the strategy consistency weight in the loss function, enabling an adaptive balance between learning new tasks and transferring knowledge from previous tasks. By generating function code offline prior to training, rather than relying on large-model inference during the training process, the SDW framework reduces computational overhead while maintaining efficiency in sequential task scenarios. Experimental results on Atari and MiniHack sequential tasks demonstrate that SDW significantly outperforms existing lifelong reinforcement learning methods.

Electronic health records (EHRs) provide a comprehensive source of longitudinal patient data, encompassing structured modalities such as laboratory results, imaging data, and vital signs, and unstructured clinical notes. These datasets, after necessary preprocessing to clean and format the data for analysis, often remain in their raw EHR form, representing numerical or categorical values without further transformation into task-agnostic embeddings. While such raw EHR data enables predictive modeling, its reliance on manual feature engineering or downstream task-specific optimization limits its utility for general-purpose applications. Deep learning (DL) techniques, such as recurrent neural networks (RNNs) and Transformers, have facilitated predictive tasks like disease progression and diagnosis prediction. However, these methods often struggle to fully exploit the temporal and multimodal dependencies inherent in EHR data due to their reliance on pre-processed but untransformed raw EHR inputs. In this study, we introduce CAAT-EHR, a novel architecture designed to bridge this gap by generating robust, task-agnostic longitudinal embeddings from raw EHR data. CAAT-EHR leverages self- and cross-attention mechanisms in its encoder to integrate temporal and contextual relationships across multiple modalities, transforming the data into enriched embeddings that capture complex dependencies. An autoregressive decoder complements the encoder by predicting future time points data during pre-training, ensuring that the resulting embeddings maintain temporal consistency and alignment. CAAT-EHR eliminates the need for manual feature engineering and enables seamless transferability across diverse downstream tasks. Extensive evaluations on benchmark datasets, demonstrate the superiority of CAAT-EHR-generated embeddings over pre-processed raw EHR data and other baseline approaches.

Accurate estimation of a user's biological joint moment from wearable sensor data is vital for improving exoskeleton control during real-world locomotor tasks. However, most state-of-the-art methods rely on deep learning techniques that necessitate extensive in-lab data collection, posing challenges in acquiring sufficient data to develop robust models. To address this challenge, we introduce a locomotor task set optimization strategy designed to identify a minimal, yet representative, set of tasks that preserves model performance while significantly reducing the data collection burden. In this optimization, we performed a cluster analysis on the dimensionally reduced biomechanical features of various cyclic and non-cyclic tasks. We identified the minimal viable clusters (i.e., tasks) to train a neural network for estimating hip joint moments and evaluated its performance. Our cross-validation analysis across subjects showed that the optimized task set-based model achieved a root mean squared error of 0.30$\pm$0.05 Nm/kg. This performance was significantly better than using only cyclic tasks (p<0.05) and was comparable to using the full set of tasks. Our results demonstrate the ability to maintain model accuracy while significantly reducing the cost associated with data collection and model training. This highlights the potential for future exoskeleton designers to leverage this strategy to minimize the data requirements for deep learning-based models in wearable robot control.