Modular AI systems can be developed using LLM-prompts-based modules to minimize deployment time even for complex tasks.However, these systems do not always perform well and improving them using the data traces collected from a deployment remains an open challenge.The data traces contain LLM inputs and outputs, but the annotations from LLMs are noisy.We hypothesize that Active Label Correction (ALC) can be use on the collected data to train smaller task-specific improved models that can replace LLM-based modules.In this paper, we study the noise in three GPT-3.5annotateddatasets and their denoising with human feedback.We also propose a novel method ALC3 that iteratively applies three updates to the training dataset: auto-correction, correction using human feedback and filtering.Our results show that ALC3 can lead to oracle performance with feedback on 17-24% fewer examples than the number of noisy examples in the dataset across three different NLP tasks.

Online learning at scale has become dramatically more popular over the last decade. While these programs provide affordable and accessible education, low retention and engagement are persistent problems. Virtual Teaching Assistants (VTAs) offer a solution: VTAs such as Jill Watson can answer questions about course logistics and content, amplifying interaction between professors and students, increasing teaching presence, and thereby improving retention and engagement. Using the Community of Inquiry framework, this paper presents what we believe is the first experimental study of the effect a VTA has on student perceptions of teaching presence, social presence, and cognitive presence. Students in a large, online, graduate computer science course were randomly assigned to sections with and without access to Jill Watson. The Community of Inquiry survey was then administered at the end of the semester to measure the three presences. We find that Jill Watson has a small, positive, statistically significant effect on the Design & Organization dimension of teaching presence as well as social presence.

Building Automatic Speech Recognition (ASR) systems for code-switched speech has recently gained renewed attention due to the widespread use of speech technologies in multilingual communities worldwide. End-to-end ASR systems are a natural modeling choice due to their ease of use and superior performance in monolingual settings. However, it is well known that end-to-end systems require large amounts of labeled speech. In this work, we investigate improving code-switched ASR in low resource settings via data augmentation using code-switched text-to-speech (TTS) synthesis. We propose two targeted techniques to effectively leverage TTS speech samples: 1) Mixup, an existing technique to create new training samples via linear interpolation of existing samples, applied to TTS and real speech samples, and 2) a new loss function, used in conjunction with TTS samples, to encourage code-switched predictions. We report significant improvements in ASR performance achieving absolute word error rate (WER) reductions of up to 5%, and measurable improvement in code switching using our proposed techniques on a Hindi-English code-switched ASR task.