Source author record

Yucheng Jiang

Yucheng Jiang appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Computation and Language cond-mat.mtrl-sci Information Retrieval physics.app-ph physics.optics

Catalog footprint

What is connected

4works

5topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Reflections and New Directions for Human-Centered Large Language Models

Large Language Models (LLMs) are increasingly shaping the private and professional lives of users, with numerous applications in business, education, finance, healthcare, law, and science. With this rise in global influence comes greater urgency to build, evaluate, and deploy these systems in a manner that prioritizes not only technical capabilities but also human priorities. This work presents a framework for developing Human-Centered Large Language Models (HCLLMs), which integrates perspectives from Natural Language Processing (NLP), Human-Computer Interaction (HCI), and responsible AI. Considering the ethics, economics, and technical objectives of language modeling, we argue that model developers need to address human concerns, preferences, values, and goals, not only during a cursory post-training stage, but rather with rigor and care at every stage of the pipeline. This paper offers human-centered insights and recommendations for developers at each stage, from system design to data sourcing, model training, evaluation, and responsible deployment. Then we conclude with a case study, applying these insights to understand the future of work with HCLLMs.

preprint2023arXiv

Effective Seed-Guided Topic Discovery by Integrating Multiple Types of Contexts

Instead of mining coherent topics from a given text corpus in a completely unsupervised manner, seed-guided topic discovery methods leverage user-provided seed words to extract distinctive and coherent topics so that the mined topics can better cater to the user's interest. To model the semantic correlation between words and seeds for discovering topic-indicative terms, existing seed-guided approaches utilize different types of context signals, such as document-level word co-occurrences, sliding window-based local contexts, and generic linguistic knowledge brought by pre-trained language models. In this work, we analyze and show empirically that each type of context information has its value and limitation in modeling word semantics under seed guidance, but combining three types of contexts (i.e., word embeddings learned from local contexts, pre-trained language model representations obtained from general-domain training, and topic-indicative sentences retrieved based on seed information) allows them to complement each other for discovering quality topics. We propose an iterative framework, SeedTopicMine, which jointly learns from the three types of contexts and gradually fuses their context signals via an ensemble ranking process. Under various sets of seeds and on multiple datasets, SeedTopicMine consistently yields more coherent and accurate topics than existing seed-guided topic discovery approaches.

preprint2020arXiv

Chargeable photoconductivity in Van der Waals heterojunctions

Van der Waals (vdW) heterojunctions, based on two-dimensional (2D) materials, show great potential for the development of eco-friendly and high-efficiency nano-devices. Considerable research has been performed and has reported valuable applications of photovoltaic cells, photodetectors, etc. However, simultaneous energy conversion and storage in a single device has not been achieved. Here, we demonstrate a simple strategy to construct a vdW p-n junction between a WSe2 layer and quasi-2D electron gas. After once optical illumination, the device stores the light-generated electrons and holes for up to seven days, and then releases a very large photocurrent of 2.9 mA with bias voltage applied in darkness; this is referred to as chargeable photoconductivity (CPC), which completely differs from any previously observed photoelectric phenomenon. In normal photoconductivity, the recombination of electron-hole pairs takes place at the end of their lifetime, causing a release of heat; in contrast, infinite-lifetime photocarriers can be generated in CPC devices without a thermal loss. The photoelectric conversion and storage are completely self-excited during the charging process. The ratio between currents in full- and empty-energy states below the critical temperature reaches as high as 109, with an external quantum efficiency of 4410000% during optical charging. A theoretical model developed to explain the mechanism of this effect is in good agreement with the experimental data. This work paves a path towards storage-type photoconductors and high-efficiency entropy-decreasing devices.

preprint2015arXiv

Raman fingerprint of semi-metal WTe2 from bulk to monolayer

Tungsten ditelluride (WTe2), a layered transition-metal dichalcogenide (TMD), has recently demonstrated an extremely large magnetoresistance effect, which is unique among TMDs. This fascinating feature seems to be correlated with its special electronic structure. Here, we report the observation of 6 Raman peaks corresponding to the A_2^4, A_1^9, A_1^8, A_1^6, A_1^5 and A_1^2 phonons, from the 33 Raman-active modes predicted for WTe2. This provides direct evidence to distinguish the space group of WTe2 from that of other TMDs. Moreover, the Raman evolution of WTe2 from bulk to monolayer is clearly revealed. It is interesting to find that the A_2^4 mode, centered at ~109.8 cm-1, is forbidden in a monolayer, which may be attributable to the transition of the point group from C2v (bulk) to C2h (monolayer). Our work characterizes all observed Raman peaks in the bulk and few-layer samples and provides a route to study the physical properties of two-dimensional WTe2.