We present systematic efforts in building long-context multilingual text representation model
(TRM) and reranker from scratch for text retrieval. We first introduce a text encoder (base …

Normalization layers have recently experienced a renaissance in the deep reinforcement
learning and continual learning literature, with several works highlighting diverse benefits …

Stochastic gradients closely relate to both optimization and generalization of deep neural
networks (DNNs). Some works attempted to explain the success of stochastic optimization …

Replacing non-polynomial functions (eg, non-linear activation functions such as ReLU) in a
neural network with their polynomial approximations is a standard practice in privacy …

Storing open-source fine-tuned models separately introduces redundancy and increases
response times in applications utilizing multiple models. Delta-parameter pruning (DPP) …

Conotoxins, being small disulfide-rich and bioactive peptides, manifest notable
pharmacological potential and find extensive applications. However, the exploration of …

Neural field methods have seen great progress in various long-standing tasks in computer
vision and computer graphics, including novel view synthesis and geometry reconstruction …

The effect of regularizers such as weight decay when training deep neural networks is not
well understood. We study the influence of weight decay as well as $ L2 $-regularization …

In continual learning settings neural network is taught different tasks sequentially and the
network is prone to catastrophic forgetting. We investigate the role of regularization methods …