Search Results for author:
Found 27 papers, 10 papers with code
Creating and Evaluating Resources for Sentiment Analysis in the Low-resource Language: Sindhi
In this paper, we develop Sindhi subjective lexicon using a merger of existing English resources: NRC lexicon, list of opinion words, SentiWordNet, Sindhi-English bilingual dictionary, and collection of Sindhi modifiers.
“Is Whole Word Masking Always Better for Chinese BERT?”: Probing on Chinese Grammatical Error Correction
The meaning of a word in Chinese is different in that a word is a compositional unit consisting of multiple characters.
Self-playing Adversarial Language Game Enhances LLM Reasoning
Hence, we are curious about whether LLMs' reasoning ability can be further enhanced by Self-Play in this Adversarial language Game (SPAG).
Prior-agnostic Multi-scale Contrastive Text-Audio Pre-training for Parallelized TTS Frontend Modeling
Furthermore, a parallelized TTS frontend model is delicately devised to execute TN, PD, and PBP prediction tasks, respectively in the second stage.
Look Before You Leap: Towards Decision-Aware and Generalizable Tool-Usage for Large Language Models
Meanwhile, we propose a novel tool sampling strategy to enhance the generalizability of LLMs over unseen tools.
Enhancing Zero-shot Counting via Language-guided Exemplar Learning
Recently, Class-Agnostic Counting (CAC) problem has garnered increasing attention owing to its intriguing generality and superior efficiency compared to Category-Specific Counting (CSC).
Embracing Language Inclusivity and Diversity in CLIP through Continual Language Learning
To alleviate CF raised by covariate shift and lexical overlap, we further propose a novel approach that ensures the identical distribution of all token embeddings during initialization and regularizes token embedding learning during training.
WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models
The rapid advancement of large language models (LLMs) has led to a new era marked by the development of autonomous applications in real-world scenarios, which drives innovation in creating advanced web agents.
Emage: Non-Autoregressive Text-to-Image Generation
Compared with autoregressive baselines that needs to run one thousand times, our model only runs 16 times to generate images of competitive quality with an order of magnitude lower inference latency.
On Diversified Preferences of Large Language Model Alignment
Our analysis reveals a correlation between the calibration performance of reward models (RMs) and the alignment performance of LLMs.
Adversarial Preference Optimization
Human preference alignment is essential to improve the interaction quality of large language models (LLMs).
TencentLLMEval: A Hierarchical Evaluation of Real-World Capabilities for Human-Aligned LLMs
We construct a hierarchical task tree encompassing 7 major areas covering over 200 categories and over 800 tasks, which covers diverse capabilities such as question answering, reasoning, multiturn dialogue, and text generation, to evaluate LLMs in a comprehensive and in-depth manner.
Everyone Deserves A Reward: Learning Customized Human Preferences
Besides, from the perspective of data efficiency, we propose a three-stage customized RM learning scheme, then empirically verify its effectiveness on both general preference datasets and our DSP set.
Chunk, Align, Select: A Simple Long-sequence Processing Method for Transformers
Although dominant in natural language processing, transformer-based models remain challenged by the task of long-sequence processing, because the computational cost of self-attention operations in transformers swells quadratically with the input sequence length.
SkillNet-X: A Multilingual Multitask Model with Sparsely Activated Skills
Traditional multitask learning methods basically can only exploit common knowledge in task- or language-wise, which lose either cross-language or cross-task knowledge.
When Federated Learning Meets Pre-trained Language Models' Parameter-Efficient Tuning Methods
To facilitate the research of PETuning in FL, we also develop a federated tuning framework FedPETuning, which allows practitioners to exploit different PETuning methods under the FL training paradigm conveniently.
Pretrained Language Encoders are Natural Tagging Frameworks for Aspect Sentiment Triplet Extraction
Existing works usually formulate the span detection as a 1D token tagging problem, and model the sentiment recognition with a 2D tagging matrix of token pairs.
Effidit: Your AI Writing Assistant
In Effidit, we significantly expand the capacities of a writing assistant by providing functions in five categories: text completion, error checking, text polishing, keywords to sentences (K2S), and cloud input methods (cloud IME).
One Model, Multiple Modalities: A Sparsely Activated Approach for Text, Sound, Image, Video and Code
Moreover, our model supports self-supervised pretraining with the same sparsely activated way, resulting in better initialized parameters for different modalities.
Pretraining Chinese BERT for Detecting Word Insertion and Deletion Errors
We achieve this by introducing a special token \texttt{[null]}, the prediction of which stands for the non-existence of a word.
MarkBERT: Marking Word Boundaries Improves Chinese BERT
We present a Chinese BERT model dubbed MarkBERT that uses word information in this work.
Chinese Named Entity Recognition
named-entity-recognition
+7
SkillNet-NLU: A Sparsely Activated Model for General-Purpose Natural Language Understanding
The key feature of our approach is that it is sparsely activated guided by predefined skills.
Exploring and Adapting Chinese GPT to Pinyin Input Method
We find that a frozen GPT achieves state-of-the-art performance on perfect pinyin.
"Is Whole Word Masking Always Better for Chinese BERT?": Probing on Chinese Grammatical Error Correction
The meaning of a word in Chinese is different in that a word is a compositional unit consisting of multiple characters.
Unsupervised Sentiment Analysis by Transferring Multi-source Knowledge
The first mechanism is a selective domain adaptation (SDA) method, which transfers knowledge from the closest source domain.
Contextualize Knowledge Bases with Transformer for End-to-end Task-Oriented Dialogue Systems
Incorporating knowledge bases (KB) into end-to-end task-oriented dialogue systems is challenging, since it requires to properly represent the entity of KB, which is associated with its KB context and dialogue context.
Ranked #2 on
Task-Oriented Dialogue Systems
on KVRET
Adversarial Training Based Multi-Source Unsupervised Domain Adaptation for Sentiment Analysis
Existing algorithms of MS-UDA either only exploit the shared features, i. e., the domain-invariant information, or based on some weak assumption in NLP, e. g., smoothness assumption.
Multi-Source Unsupervised Domain Adaptation
Sentiment Analysis
+2
