Comparing Fine-Tuning, Zero and Few-Shot Strategies with Large Language Models in Hate Speech Detection in English

Large Language Models(LLMs)are increasingly demonstrating their ability to understand natural language and solve complex tasks,especially through text generation.One of the relevant capabilities is contextual learning,which involves the ability to receive instructions in natural language or task demonstrations to generate expected outputs for test instances without the need for additional training or gradient updates.In recent years,the popularity of social networking has provided a medium through which some users can engage in offensive and harmful online behavior.In this study,we investigate the ability of different LLMs,ranging from zero-shot and few-shot learning to fine-tuning.Our experiments show that LLMs can identify sexist and hateful online texts using zero-shot and few-shot approaches through information retrieval.Furthermore,it is found that the encoder-decoder model called Zephyr achieves the best results with the fine-tuning approach,scoring 86.811%on the Explainable Detection of Online Sexism(EDOS)test-set and 57.453%on the Multilingual Detection of Hate Speech Against Immigrants and Women in Twitter(HatEval)test-set.Finally,it is confirmed that the evaluated models perform well in hate text detection,as they beat the best result in the HatEval task leaderboard.The error analysis shows that contextual learning had difficulty distinguishing between types of hate speech and figurative language.However,the fine-tuned approach tends to produce many false positives.

An Adaptive Hate Speech Detection Approach Using Neutrosophic Neural Networks for Social Media Forensics

Detecting hate speech automatically in social media forensics has emerged as a highly challenging task due tothe complex nature of language used in such platforms. Currently, several methods exist for classifying hatespeech, but they still suffer from ambiguity when differentiating between hateful and offensive content and theyalso lack accuracy. The work suggested in this paper uses a combination of the Whale Optimization Algorithm(WOA) and Particle Swarm Optimization (PSO) to adjust the weights of two Multi-Layer Perceptron (MLPs)for neutrosophic sets classification. During the training process of the MLP, the WOA is employed to exploreand determine the optimal set of weights. The PSO algorithm adjusts the weights to optimize the performanceof the MLP as fine-tuning. Additionally, in this approach, two separate MLP models are employed. One MLPis dedicated to predicting degrees of truth membership, while the other MLP focuses on predicting degrees offalse membership. The difference between these memberships quantifies uncertainty, indicating the degree ofindeterminacy in predictions. The experimental results indicate the superior performance of our model comparedto previous work when evaluated on the Davidson dataset.

Chaotic Elephant Herd Optimization with Machine Learning for Arabic Hate Speech Detection

In recent years,the usage of social networking sites has considerably increased in the Arab world.It has empowered individuals to express their opinions,especially in politics.Furthermore,various organizations that operate in the Arab countries have embraced social media in their day-to-day business activities at different scales.This is attributed to business owners’understanding of social media’s importance for business development.However,the Arabic morphology is too complicated to understand due to the availability of nearly 10,000 roots and more than 900 patterns that act as the basis for verbs and nouns.Hate speech over online social networking sites turns out to be a worldwide issue that reduces the cohesion of civil societies.In this background,the current study develops a Chaotic Elephant Herd Optimization with Machine Learning for Hate Speech Detection(CEHOML-HSD)model in the context of the Arabic language.The presented CEHOML-HSD model majorly concentrates on identifying and categorising the Arabic text into hate speech and normal.To attain this,the CEHOML-HSD model follows different sub-processes as discussed herewith.At the initial stage,the CEHOML-HSD model undergoes data pre-processing with the help of the TF-IDF vectorizer.Secondly,the Support Vector Machine(SVM)model is utilized to detect and classify the hate speech texts made in the Arabic language.Lastly,the CEHO approach is employed for fine-tuning the parameters involved in SVM.This CEHO approach is developed by combining the chaotic functions with the classical EHO algorithm.The design of the CEHO algorithm for parameter tuning shows the novelty of the work.A widespread experimental analysis was executed to validate the enhanced performance of the proposed CEHOML-HSD approach.The comparative study outcomes established the supremacy of the proposed CEHOML-HSD model over other approaches.

Hate speech detection in Twitter using hybrid embeddings and improved cuckoo search-based neural networks

Purpose-Hate speech is an expression of intense hatred.Twitter has become a popular analytical tool for the prediction and monitoring of abusive behaviors.Hate speech detection with social media data has witnessed special research attention in recent studies,hence,the need to design a generic metadata architecture and efficient feature extraction technique to enhance hate speech detection.Design/methodology/approach-This study proposes a hybrid embeddings enhanced with a topic inference method and an improved cuckoo search neural network for hate speech detection in Twitter data.The proposed method uses a hybrid embeddings technique that includes Term Frequency-Inverse Document Frequency(TF-IDF)for word-level feature extraction and Long Short Term Memory(LSTM)which is a variant of recurrent neural networks architecture for sentence-level feature extraction.The extracted features from the hybrid embeddings then serve as input into the improved cuckoo search neural network for the prediction of a tweet as hate speech,offensive language or neither.Findings-The proposed method showed better results when tested on the collected Twitter datasets compared to other related methods.In order to validate the performances of the proposed method,t-test and post hoc multiple comparisons were used to compare the significance and means of the proposed method with other related methods for hate speech detection.Furthermore,Paired Sample t-Test was also conducted to validate the performances of the proposed method with other related methods.Research limitations/implications-Finally,the evaluation results showed that the proposed method outperforms other related methods with mean F1-score of 91.3.Originality/value-The main novelty of this study is the use of an automatic topic spotting measure based on na€ıve Bayes model to improve features representation.

Annoyance-type speech emotion detection in working environment

In order to recognize people＇s annoyance emotions in the working environment and evaluate emotional well- being, emotional speech in a work environment is induced to obtain adequate samples of emotional speech, and a Mandarin database with two thousands samples is built. In searching for annoyance-type emotion features, the prosodic feature and the voice quality feature parameters of the emotional statements are extracted first. Then an improved back propagation （BP） neural network based on the shuffled frog leaping algorithm （SFLA） is proposed to recognize the emotion. The recognition capability of the BP, radical basis function （RBF） and the SFLA neural networks are compared experimentally. The results show that the recognition ratio of the SFLA neural network is 4. 7% better than that of the BP neural network and 4. 3% better than that of the RBF neural network. The experimental results demonstrate that the random initial data trained by the SFLA can optimize the connection weights and thresholds of the neural network, speed up the convergence and improve the recognition rate.

Modified Cepstral Feature for Speech Anti-spoofing

The hidden danger of the automatic speaker verification(ASV)system is various spoofed speeches.These threats can be classified into two categories,namely logical access(LA)and physical access(PA).To improve identification capability of spoofed speech detection,this paper considers the research on features.Firstly,following the idea of modifying the constant-Q-based features,this work considered adding variance or mean to the constant-Q-based cepstral domain to obtain good performance.Secondly,linear frequency cepstral coefficients(LFCCs)performed comparably with constant-Q-based features.Finally,we proposed linear frequency variance-based cepstral coefficients(LVCCs)and linear frequency mean-based cepstral coefficients(LMCCs)for identification of speech spoofing.LVCCs and LMCCs could be attained by adding the frame variance or the mean to the log magnitude spectrum based on LFCC features.The proposed novel features were evaluated on ASVspoof 2019 datase.The experimental results show that compared with known hand-crafted features,LVCCs and LMCCs are more effective in resisting spoofed speech attack.

Speech Signal Detection Based on Bayesian Estimation by Observing Air-Conducted Speech under Existence of Surrounding Noise with the Aid of Bone-Conducted Speech

In order to apply speech recognition systems to actual circumstances such as inspection and maintenance operations in industrial factories to recording and reporting routines at construction sites, etc. where hand-writing is difficult, some countermeasure methods for surrounding noise are indispensable. In this study, a signal detection method to remove the noise for actual speech signals is proposed by using Bayesian estimation with the aid of bone-conducted speech. More specifically, by introducing Bayes’ theorem based on the observation of air-conducted speech contaminated by surrounding background noise, a new type of algorithm for noise removal is theoretically derived. In the proposed speech detection method, bone-conducted speech is utilized in order to obtain precise estimation for speech signals. The effectiveness of the proposed method is experimentally confirmed by applying it to air- and bone-conducted speeches measured in real environment under the existence of surrounding background noise.

Automated Speech Recognition System to Detect Babies’ Feelings through Feature Analysis

Diagnosing a baby’s feelings poses a challenge for both doctors and parents because babies cannot explain their feelings through expression or speech.Understanding the emotions of babies and their associated expressions during different sensations such as hunger,pain,etc.,is a complicated task.In infancy,all communication and feelings are propagated through cryspeech,which is a natural phenomenon.Several clinical methods can be used to diagnose a baby’s diseases,but nonclinical methods of diagnosing a baby’s feelings are lacking.As such,in this study,we aimed to identify babies’feelings and emotions through their cry using a nonclinical method.Changes in the cry sound can be identified using our method and used to assess the baby’s feelings.We considered the frequency of the cries from the energy of the sound.The feelings represented by the infant’s cry are judged to represent certain sensations expressed by the child using the optimal frequency of the recognition of a real-world audio sound.We used machine learning and artificial intelligence to distinguish cry tones in real time through feature analysis.The experimental group consisted of 50%each male and female babies,and we determined the relevancy of the results against different parameters.This application produced real-time results after recognizing a child’s cry sounds.The novelty of our work is that we,for the first time,successfully derived the feelings of young children through the cry-speech of the child,showing promise for end-user applications.

A recursive calculating algorithm for higher-order cumulants over sliding window and its application in speech endpoint detection

Regarding the performance of traditional endpoint detection algorithms degrades as the environment noise level increases, a recursive calculating algorithm for higher-order cu- mulants over a sliding window is proposed. Then it is applied to the speech endpoint detection. Furthermore, endpoint detection is carried out with the feature of energy. Experimental results show that both the computational efficiency and the robustness against noise of the proposed algorithm are improved remarkably compared with traditional algorithm. The average prob- ability of correct point detection （Pc-point） of the proposed voice activity detection （VAD） is 6.07% higher than that of G.729b VAD in different noisy at different signal-noise ratios （SNRs） environments.

Speech endpoint detection in low-SNRs environment based on perception spectrogram structure boundary parameter

The Perception Spectrogram Structure Boundary（PSSB）parameter is proposed for speech endpoint detection as a preprocess of speech or speaker recognition.At first a hearing perception speech enhancement is carried out.Then the two-dimensional enhancement is performed upon the sound spectrogram according to the difference between the determinacy distribution characteristic of speech and the random distribution characteristic of noise.Finally a decision for endpoint was made by the PSSB parameter.Experimental results show that,in a low SNR environment from-10 dB to 10 dB,the algorithm proposed in this paper may achieve higher accuracy than the extant endpoint detection algorithms.The detection accuracy of 75.2%can be reached even in the extremely low SNR at-10 dB.Therefore it is suitable for speech endpoint detection in low-SNRs environment.

Speech Recognition for Parkinson’s Disease Based on Improved Genetic Algorithm and Data Enhancement Technology

Parkinson’s disease is one of the most destructive diseases to the nervous system.Speech disorder is one of the typical symptoms of Parkinson’s disease.Approximately 90%of Parkin-son’s patients develop some degree of speech disorder,which affects speech function faster than any other subsystem of the body.Screening Parkinson’s disease by sound is a very effective method that has attracted a growing number of researchers over the past decade.Patients with Parkinson’s disease could be identified by recording the sound signal of the pronunciation of words,extracting appropriate features and identifying the disturbance in their voices.This paper proposes an improved genetic algorithm combined with a data enhancement method for Parkinson’s speech signal recognition.Specifically,the methods first extract representative speech signal features through the L1 regularization SVM and then enhance the representative feature data by the SMOTE algorithm.Following this,both original and enhanced features are used to train an SVM classifier for speech signal recognition.An improved genetic algorithm was applied to find the optimal parameters of the SVM.The effectiveness of our proposed model is demonstrated by using Parkinson’s disease audio data set from the UCI machine learning library,and compared with the most advancedmethods,our proposed method has the best performance.