About this document
This is an English translation of a crowdsourced document created by Chinese PhD students to share observations about advisor experiences, lab environments, and advising styles. This document includes both negative reports ("red flags") and rebuttals from lab members defending their advisors. Many entries are disputed. Please read critically and form your own judgment. The original document explicitly warns against fabricating information and encourages factual, responsible sharing.
⚠️ Important: Some entries in this document are actively contested. Tags marked "Disputed" have received rebuttals from people claiming the negative reports are false or exaggerated. This translation preserves both critical accounts and rebuttals for balanced reading.
🏛 Stanford University
Stanley Lei Qi
Students and postdocs reportedly suffer systematic psychological pressure and depression. Accused of sexual harassment of female students at events, favoritism toward connected families, and blocking graduation for underperformers.
Multiple PhD students and postdocs have reportedly suffered serious psychological pressure (PUA-style manipulation) and varying degrees of depression. Students are treated as disposable — those who don't produce results face constant coldness and are blocked from graduating.
Allegedly sexually harasses female students at events. Accused of operating a clique (particularly with Tsinghua connections), suppressing students, and engaging in favoritism/nepotism (powerful families reportedly place children in his lab).
Jiajun Wu
Isolated reports of students struggling or receiving negative references; strongly rebutted by many others who had excellent experiences. High competition among interns may affect outcomes relative to expectations.
Reports of students returning from summer research in poor mental states — one reportedly went into industry despite having multiple first-author papers; another allegedly received a negative recommendation letter.
Rebuttal from lab members
Multiple people who did summer research in Jiajun's lab report very positive experiences: fastest period of research growth, very nice advisor, all interns from one cohort got into top-10 PhD programs. Jiajun is described as extremely supportive for PhD applications — proactively reaching out to discuss applications, submitting 15+ letters within 5 minutes. Caveats: the lab is competitive with many strong interns, so outcomes relative to expectations may vary.
Shuran Song
Vague references to a past controversy; some say behavior changed significantly after tenure. Current student experience appears positive. Limited concrete details available.
One comment says Shuran is "a very excellent professor" and that the difference in student happiness between her lab and Fisher Yu's lab shows who caused a past controversy. Another notes she treats students well now but was reportedly different before tenure.
🏛 Carnegie Mellon University (CMU)
Min Xu
Runs an enormous "fishing pond" of 1,000+ interns with virtually no individual guidance. Low-quality postdocs do all supervision. Lab severely underfunded (100+ people, one GPU). Rarely converts to PhD offers despite promises; known for years of PhD students quitting.
Reportedly has 1,000+ interns in a Slack group with 2,000+ people (free plan). Classic "fishing pond" — recruits heavily but provides no real guidance.
Postdocs are reportedly low quality — some can't read a loss function — and provide no mentorship. Only pushes students without giving direction.
Lab is reportedly very underfunded — 100+ people sharing one 3090 GPU server. Advertises grand plans to Research Assistants but rarely converts to PhD slots. Known for years of not attracting new PhD students; previous ones reportedly quit.
Considered one of the "three high-volume paper producers" in Chinese CS academia (alongside Dacheng Tao and Philip S. Yu).
Justin Chan
Dismissed 2 of 3 PhD students within his first year. Made sexist remarks about female researchers. Allegedly stole a visiting PhD student's 6 months of work and gave credit to his own student.
Recruited 3 PhD students in his first year as AP; dismissed 2 within less than a year.
Allegedly said: "Female researchers always have to rely on male researchers."
Reportedly took credit from a visiting PhD student: the visiting student worked 6 months on a project; when his own PhD student couldn't produce results, he gave the visiting student's work to his own PhD student, who ended up as 5th author while the visiting student got nothing.
🏛 Rice University
Tong (Tony) Geng
All 3 first-year students at Rochester quit; more followed at Rice. Accused of threatening to block degrees, requiring secret NDAs, late-night calls, and sabotaging students who tried to switch to other advisors. More recent accounts suggest improvement.
At his previous position at University of Rochester ECE (as AP), all 3 of his first-year students quit before the year ended. He was also reported to the department and college. The story reportedly became well-known across all of Rochester. He then moved to Rice, where more students reportedly quit.
Lab homepage membership changes frequently — students come and go rapidly.
One person's close friend was in the lab and showed significant mental health deterioration — stopped socializing and playing sports.
Reportedly required students to sign confidentiality agreements privately, prohibiting sharing of any work including ideas, slides, and papers — threatened jail for violations.
Reportedly said to a student wanting to withdraw: "If you quit, I'll resign! Let's see if the department chair listens to you or me! You won't even get a Master's degree! (The chair and I are buddies!)"
Allegedly required an older student to "donate" a first-authorship to each new student as a "tradition" — this was reportedly reported to the department chair.
Late-night phone calls (reportedly 1am) to students.
Rebuttal / Other Perspective
One account says the first-year situation was due to a new AP being too aggressive under work pressure — students who quit did publish decent papers and found good positions afterward. A more recent account from a friend in the lab says things are now normal: professor is professional, doesn't contact after 6pm or on weekends, communicates via Teams/email not WeChat, keeps promises, and gives hands-on guidance.
🏛 UC Riverside
Zheng Sika
Systematic pattern of blocking student progress: gives failing grades to students who try to switch advisors, drags projects for years, disappears for months then demands full redos, and failed a student in their final month before graduation. Requires daily and weekly reports with 24-hour turnaround.
Has funding but keeps students as TAs; when asked about becoming an RA, response is "you're not working hard enough."
Can drag a single rotation project on for 4 years.
When students try to switch advisors, reportedly gives them a failing grade without hesitation — exploiting the school policy that 2 F's result in automatic dismissal.
Disappears for 2 months (vacation), then returns and requires a complete redo of all work after a 30-minute review.
Two students switched labs / dropped out within one year; one persistent student was failed right before their final month of graduation.
Wrote in an annual progress review: "None of their stuff makes any sense / I don't endorse it."
Requires daily reports; weekly reports must be revised within 24 hours; if the report is on a Friday, the entire weekend is consumed by corrections.
Told a student the week before their candidacy exam to redo everything, and when asked what to do, replied: "Do your candidacy exam."
Attempted to claim authorship on an unrelated paper from a student who switched labs, with threats of legal action without evidence.
🏛 Northwestern University
Kaize Ding
Disputed allegations of an inappropriate relationship with a female student; the student who filed a complaint was later dismissed reportedly for academic misconduct. Many interns and collaborators report strongly positive experiences — helpful, fair, responsive.
Controversy with a female student publicized on Chinese social media (Xiaohongshu). Allegations of an inappropriate relationship. The student who reportedly filed a complaint was later dismissed — but reportedly for academic misconduct, with the person then allegedly posting defamatory content.
Criticism of research output: accused of "pumping out low-quality GNN papers." Counter-arguments note the field itself has matured and he now supports newer directions (agents, LLMs).
Rebuttal from former intern
Former intern says Kaize is excellent to students and interns: very actively helps find internships (personally used his connections), writes thorough recommendation letters (helped with ~20 applications), meticulously reviews papers, fair authorship (by contribution). Multiple people strongly endorse him as a good advisor.
🏛 Indiana University Bloomington
Fengguang Song
Uses ChatGPT to respond to students and forces research based on outdated AI-found papers. Weaponizes administrative tools: blocks credits, withholds scholarships, refuses internship paperwork, then suddenly demands students quit. Five students left within 2 years of him taking over the lab.
After taking over a lab from a previous professor, 5 students left within 2 years.
Weak research ability — reportedly uses ChatGPT to respond to students, including live during lab meetings.
Forces students to read 5–10 year old papers found via AI search, treating them as ground truth.
Teaches his own online course but gives his lab's PhD students B grades while tightly monitoring attendance.
Sends large batches of AI-generated emails every Saturday night at 11pm.
Uses administrative tools to make life difficult: goes back on promises about internships; deliberately blocks required credits; manipulates scholarship disbursement.
Refused to sign a CPT letter for a student's fall internship (citing "prepare to graduate"), then on the first week of fall semester told the student to find a new advisor or quit.
🏛 MIT
Wojciech Matusik
Mild criticism about one prolific lab member publishing many papers — largely rebutted by lab members who report a positive, well-funded environment. The specific authorship concern was clarified and accepted by the person involved.
Some criticism about a lab member (Zhiyang Duo) publishing many papers with 10+ interns, 11 affiliations on one paper. The original critic retracted, clarifying it was a dataset paper where many students reviewed data quality, justifying the author list.
Rebuttal from lab member
Lab atmosphere described as good. Matusik himself is described as extremely hardworking. Good funding — free lunch/dinner, large refrigerator. The domain is actually difficult to publish in, making "high volume" criticism less fair.
Manya Ghobadi
Accused of fabricating data, having limited domain knowledge, and maintaining her position through her partner (Adam Chlipala, MIT). Gives undergraduates first authorship over grad students; dismissed a Chinese postdoc with visa complications. Severely underfunded, using a startup to cover expenses.
Reportedly left Google in under a year. Went to Microsoft and took a step back in role (postdoc).
MIT faculty position allegedly obtained and maintained through her partner (Adam Chlipala, MIT faculty).
Allegedly fabricates data in papers; focuses on storytelling rather than new algorithms.
Does not actually understand networking, security, or optics — but insists on working in optical network security.
Reportedly has emotional breakdowns in 1-on-1 meetings and treats students as therapists.
Reportedly keeps students waiting 30+ minutes outside her office and blames being busy.
Authorship reportedly unfair — graduate students who do the most work must yield first authorship to undergraduates to make her look like a great mentor.
Dismissed a male Chinese postdoc, leaving him in J-visa difficulty; he was absorbed by her partner's group.
Reportedly severely underfunded; uses a startup to keep appearances and pressures students to work there.
🏛 NJIT
Hai Phan
Accused of mocking a candidate's accent during an interview and dismissing their research arrogantly. Single account — limited information available.
Accused of racial discrimination — reportedly mocked an applicant's accent during an interview and repeatedly questioned the value of their research in an arrogant manner.
🏛 UC San Diego
Zhiting Hu
Went to xAI mid-advising. Accused of being part of a 'couples shop' and weak academically — all strongly rebutted. Lab is described as extremely well-resourced with fast PhD completions and excellent student outcomes.
Left UCSD for xAI; unclear what happens to remaining PhD students.
Accused of being a "couple shop" (spouses Zhiting Hu and Lianhui Qin sharing PhD students).
Some claim he's weak academically and "relies on his wife."
Rebuttal
Countered strongly: most PhDs graduated on time (3–4 years), good industry funding, excellent student outcomes (top-10 faculty positions, top companies). Students work in industry part-time using industry resources — described as appropriate for the field. Described as a "rising star" with many awards. Wrote strong recommendation letters for all interns. Lab is said to be well-resourced (one of the best-funded in the US).
Lianhui Qin
Minor criticism: tends to ghost meetings. Widely regarded as a good advisor overall. Sharing PhD students with her partner Zhiting Hu is noted but not considered a serious problem.
Accused of being weak academically and "relying on her husband" (Zhiting Hu). Counter-comments suggest she tends to ghost meetings but is otherwise fine. Many say she is a good advisor overall.
Nuno Vasconcelos
Average time-to-degree is reportedly 7 years; requires 7+ top-conference first-author papers before allowing a defense. Poor industry relationships isolate students. Pits two new PhD students against each other for the single funding slot.
Described as a "legendary graduation delayer" — average time-to-degree reportedly 7 years.
Reportedly requires 7+ top-conference first-author papers before allowing a defense.
Reportedly out of touch with current research; uses 8-GPU clusters to do pretraining.
Poor industry relationships — students struggle to find internships and some are on industry blacklists.
Recruits 2 first-year PhD students and lets them compete for funding; whoever produces results first gets funded, the other must switch or quit. (Note: reportedly due to funding shortfalls, not deliberate "gladiator" strategy.)
Xinyu Zhang
Known workaholic who may neglect personal health and family — not directly harmful to students. One former student praises him as a strong mentor who provides excellent research training.
Described as a "grind culture" professor — workaholic who neglects family and health, very push-oriented.
Rebuttal
A former MS student says Zhang is a good person, taught a lot, was likely very helpful for PhD applications, and that following him seriously leads to great outcomes. His being a workaholic is noted as genuine, not performative.
🏛 UT Dallas
Xinya Du
Repeatedly late to meetings without notice; unfocused during meetings; forces sick students to work. Multiple students have quit. Communication reported as consistently unclear. Attempted multiple times to move institutions without success.
Meeting frequently late without notice; during meetings, spends 50%+ of the time asking the same questions repeatedly and misses the research focus.
Forces students to work when ill.
Communication problems — unclear in both Chinese and English.
Multiple students have quit (described as a red flag to avoid).
Reportedly made multiple failed attempts to move to a better institution.
🏛 UC Santa Cruz
Chenguang Wang
All first-year PhD students quit; he then allegedly spread rumors about them. Forced students to cancel rotations in his presence and email other advisors with him BCCed. Sent legal threats to someone who posted a public warning. Cancels submissions hours before deadlines, bans internal collaboration, forced a Master's student to graduate late. Opened Zoom to yell at students who left after 3am.
Described as extremely inconsistent — reverses decisions frequently (e.g., agrees to submit to a conference then cancels hours before the deadline).
Sent a legal threat letter (via a Chinese law firm) to someone who posted a Xiaohongshu warning about him.
All first-year PhD students reportedly quit, and he then allegedly spread rumors about those who left.
Refuses to let lab members collaborate with each other; ignores warnings about scope overlap with competing labs.
Threatened a Master's student's graduation — reportedly followed through, delaying graduation by a semester.
Required a student to write an email in front of him canceling their rotation at another lab — with that email BCC'd to him.
Forced students to work until 3am; opened Zoom calls and yelled at students who left early.
Moved to UCSC from WUSTL after reportedly failing to keep students (several were reportedly taken away by other advisors).
🏛 University of Chicago
Junchen Jiang
Only vague concern raised about lab atmosphere — no specific details provided. Treat with low confidence.
Lab atmosphere mentioned — no specific details provided in the original document beyond vague concerns.
🏛 Princeton University
Felix Heide
Described as Princeton CS's most toxic professor — threatens to block graduation, internships, qualifying exams, and funding. Reported to the department dozens of times with no change. All early PhD students reportedly left. Poor reputation within Stanford and the broader field.
Described as "the most outrageous professor I've ever seen" and "the most toxic in Princeton CS, bar none."
Extreme ego; reportedly threatens to block student graduation, internships, qualifying exams, and funding.
Reported to the department dozens of times — but continues the same behavior.
Poor reputation in academia and industry; reportedly bad relationship with Stanford colleagues.
Reportedly all early PhD students either got master's degrees and left, quit, or transferred.
Mengdi Wang
Recruits many interns; a postdoc in her group has a reputation for authorship disputes. Allegedly doesn't fund everyone after over-recruiting. Collaborators say she is personally kind and busy.
Recruits very many interns — described as a "couples shop" (married to Prof. Congle from TAMU).
A postdoc in the group (Northwestern background, Peking University PhD) reportedly habitually steals first authorship — well-known for this on Xiaohongshu.
Allegedly doesn't fund people after over-recruiting.
Rebuttal
One collaborator says Prof. Wang is very nice and very busy, and the experience was positive overall.
Zhuang Liu
A student reportedly quit or switched in his first AP year. Described as treating interns poorly with no return offers. Suspected of manipulating author order on papers.
Student reportedly quit (or switched advisors) in his first year as AP.
Reportedly treats interns poorly; no return offers given to any. Counter-claim: a few interns had good PhD application outcomes despite no return.
Suspicion of academic integrity issues — author order allegedly manipulated on papers.
🏛 University of Pennsylvania
Lingjie Liu
Accusations of poor skills, controlling behavior, and verbal attacks on students in meetings — all heavily disputed. Multiple collaborators and interns report very positive experiences. MPI dismissal story specifically denied by a witness. Married to Jiatao Gu (UPenn), who is widely praised.
Accused of having poor technical skills, high control, being mean to students, and verbally attacking students in meetings.
2 students out of 5 left — but claimed they were dismissed, not that they quit; one reportedly transferred to an NLP lab; the other re-applied to PhD programs.
Reportedly poor reputation makes it hard to recruit good students.
Allegedly threatened students when she was a postdoc at MPI.
Authorship of key papers questioned.
Rebuttal from collaborators and others
Multiple collaborators say she is very reliable, willing to provide resources, has very broad connections, deep understanding of vision topics, and was very helpful. Former intern says she is kind and approachable. An MPI person who was there at the time says the dismissal story is incorrect — the student was struggling and better suited for HCI; the decision came from the department, not Lingjie. One collaborator specifically praised her as among the best they've worked with for paper feedback.
Jiatao Gu
Uniformly positive mentions — full of research insight, hard-working but not coercive. Apple background. Married to Lingjie Liu (UPenn). Widely recommended as an excellent advisor.
Generally positive
Multiple people say Jiatao is very good — full of insights, very nice, very hard-working but doesn't pressure students. Married to Lingjie Liu (UPenn). Came from Apple; described as a great advisor who should be on the "positive" list. Several people strongly recommend him.
🏛 Columbia University
Sharon Di
Recruits dozens of MS students in a school-wide competition — only the best performer gets a PhD spot. Extremely demanding with regular weekend and holiday meetings as the norm.
Recruits dozens of MS students school-wide as a "gladiator pit" — the best performer gets to stay for a PhD.
Extremely demanding — meetings on weekends, evenings, and holidays are normal.
Xiaodong Wang
A student reportedly developed depression in the lab. Accused of running paid external projects and directing student labor toward them.
Student reportedly developed depression.
Allegedly runs external paid projects.
Baron Law
Adjunct who funnels students into his company with no computing resources or research direction. Students developed depression; he then micromanaged and monitored attendance. Openly made discriminatory remarks against mainland Chinese students. Eventually assigned students to tasks completely unrelated to research.
Adjunct professor who recruits students to intern at his company — but provides no computing resources or research direction.
Students reportedly developed depression; he then engaged in office politics and micromanagement.
Monitors attendance and working hours closely; eventually assigned students tasks completely unrelated to research.
Reportedly a Hong Kong native who openly discriminates against mainland Chinese students.
🏛 UC Berkeley
Sergey Levine
A Reddit post about lab abuse at BAIR was linked, though the bad actor was reportedly a postdoc, not Levine directly. Accused of forcing students into unwanted research directions and running intern competitions where only the winner gets strong support. Over 90% of papers reportedly cannot be reproduced.
A Reddit post about abuse at BAIR was linked — the abuse reportedly came primarily from a postdoc, not Levine directly.
Reportedly forces students into research directions they don't like.
Over 90% of papers reportedly cannot be reproduced.
Reportedly runs many intern competitions with only the "winner" getting a strong recommendation.
🏛 Johns Hopkins University
Mathias Unberath
Recruits many interns, doesn't know their names, promises first-authorship then redirects it to junior PhD students. Refuses or provides negative recommendation letters for interns. Essentially uses interns as free labour to produce results for his PhD students, who are treated better.
Very bad temper; recruits many interns but doesn't know them personally (can't even recognize their names).
Promises interns first-authorship opportunities after doing "horizontal tasks" — but when work is completed, the first authorship goes to a junior PhD student, second authorship to another lab member, and extra authors are added without contribution.
Refuses to write recommendation letters for interns, or writes negative ones.
Reportedly treats his own PhD students better than interns; essentially exploits interns to produce results for PhDs.
Jason Eisner
Exceptionally strict academic standards — described as an outstanding advisor only suitable for extremely capable and resilient students. Not harmful, just intensely demanding.
Extremely strict. Described as "definitely an excellent advisor" but requiring absolute confidence in your intelligence and physical stamina.
🏛 University of Illinois Urbana-Champaign
Jimeng Sun
Accused of dismissing all students when moving institutions without notice, and of pressuring a student who later died to submit a paper right up to their death. Allegedly discriminates against women and Chinese students while favouring white students. Abandoned family for a partner and reportedly directs students toward her company.
When moving from Georgia Tech to UIUC, reportedly dismissed essentially all his existing students without proper arrangements.
A student reportedly died before a WWW 2021 paper deadline — Jimeng allegedly insisted they submit the paper regardless ("even if you have to fabricate data"), and the student died. This is stated as known within the data mining community.
At Georgia Tech: Chinese students without a first-author paper in year 1 were reportedly dismissed; students in their 3rd or 4th year were also dismissed.
Allegedly abandoned his family for another relationship, and reportedly directs students to work at his partner's company.
Accused of discriminating against women and favoring white students over pushing Chinese students.
🏛 UW-Madison
Sikai Chen
Extreme personal exploitation: forced students to drive him to the airport at night, do dishes at his home, and pay his rent when he was a postdoc. Bought his wife a computer instead of his student's; expensed his wife's meals but not student meetings. Accused of fabricating papers as a student, stealing authorship, and requiring all emails to be CCed to him. Technically described as clueless.
Accused of fabricating papers when he was a student.
Pushed a student to write proposals while they were recovering from back surgery (bedridden).
Forced students to drive him to the airport in the middle of the night; required students to do dishes at his home.
Did not buy student computers but bought one for his wife; doesn't reimburse meeting expenses for students but does for his wife's meals.
When he was a postdoc, reportedly made students pay his rent.
Steals junior students' first authorship; requires all labmate-to-advisor emails to CC him.
Invites students to dinner, then makes students pay.
Technically clueless — reportedly pointed at unsupervised learning and asked "Is this supervised or what? I don't know, do you?"
Sharon Li
As an AP, reportedly cut funding to students without papers within 6 months. After tenure, only 3 graduates, two reportedly estranged from her. Allegedly withdrew RA offers after making them. Described as "mean" by multiple people, though at least one dismissal claim is directly disputed.
As an AP, allegedly cut funding and forced students to switch labs if no paper was produced within 6 months; many students quit or switched. After tenure, only 3 students had graduated, two of whom reportedly had falling-outs with her.
Allegedly withdrew RA offers after making them — leaving students without funding.
Described as "mean" by multiple people.
Rebuttal
One person disputes the "falling out" claim, saying: "Not true — there was no falling out."
Xiaobin Xiong
Left for China mid-way, leaving PhD students in limbo. Banned students from submitting to CoRL. Research direction described as disconnected from AI despite 'AI lab' branding. Consistently rated 1-star on Rate My Professor.
Returned to China (Shanghai Chuangzhi); lab direction branded as "Legged AI Lab" despite allegedly never working in AI.
Called CoRL a "low-quality conference" and prohibited students from submitting there.
Return to China left PhD students without a clear supervisor.
Research ideas described as "dreamlike" — technically weak, especially in learning-based methods.
Rate My Professor: consistently 1 star.
Wenxiao Pan
Collectively reported to the department by lab students. Very limited public information available.
Reportedly was collectively reported to the department by students in the lab.
🏛 Texas A&M University
Yue Zhang (TAMU)
Accused of discriminating against Asian and female students while favouring white students. Every student in the lab reportedly has conflicts with him; some have needed therapy or had extreme reactions. Described as academically weak, relying on a postdoc collaborator for major results. Spreads rumours about other PIs to students.
Chinese professor but reportedly discriminates against Asian students while favoring white students; also accused of gender discrimination against women.
Multiple postdocs, PhDs, and even undergraduates have quit. Every student still in the lab reportedly has conflicts with him.
Reportedly exhibits NPD (Narcissistic Personality Disorder) traits — gaslighting students to the point where they need therapy; at least one student had more extreme reactions.
Weak research background — reportedly was pushed to graduate by his own PhD advisor. Current major papers rely on a postdoc collaborator.
Talks badly about and spreads rumors about other PIs to students.
Private life reportedly chaotic — even his friends are distressed by the situation.
🏛 Georgia Tech
Yingyan (Celine) Lin
Five or more PhD students quit or transferred in two years. Accused of NPD-type behaviour, internal student voting to expel peers, and recycling a few ideas across many papers. One student defends her funding and mentorship support.
Reportedly ghosts interview applicants.
5+ PhD students quit or transferred in the past 2 years.
Good to early graduates; quality deteriorated for later cohorts.
Described as academically mediocre; early success reportedly due to guidance from a collaborator (Atlas Wang at TAMU) rather than her own direction.
Exhibits NPD-type behavior; lab members seek therapy.
Reportedly had students vote to expel a fellow student — described as eyebrow-raising behavior.
Publishes similar ideas recycled across 3–4 different papers.
Rebuttal from a PhD student in the lab
"Prof. Lin funds students, helps revise papers, teaches presentation skills, and helps find internships/jobs. Don't let bad-faith actors mislead you."
🏛 UC Merced
Yiwei Wang
Very strongly disputed. Dozens of current and former students and collaborators wrote lengthy rebuttals praising hands-on guidance, self-funded GPUs, late-night emotional support, and fair authorship. Original complaints appear isolated and are directly contradicted by a large body of positive accounts.
Original criticism: too many interns; provides no real guidance; reportedly dismissive in interviews.
Graduation requirements listed as: 7 CCF-A first-author papers = 3-year graduation; 6 papers = 3.5 years; 5 papers = 4 years. All must be sole first-authored by the student with the professor as sole corresponding author.
Large volume of arXiv preprints (~90 listed, counter-claim: fewer than 90 actual) — code and benchmarks not open-sourced.
Accused of writing negative recommendation letters for some students.
"Couples shop" with collaborator Yujun Cai (University of Queensland, Australia).
Very strong rebuttal from multiple students and collaborators
Numerous people — students, former interns, and collaborators — wrote strongly positive accounts. Highlights include: helps students with absolutely no prior publications get their first A-conference paper; personally edits figures, rewrites sentences, brainstorms paper titles; available via WeChat and responds promptly; gives ideas proactively; provides 8x A100 GPUs (self-funded); focuses on student wellbeing and development; fair authorship. His collaborator Yujun Cai is also described very positively: edits word by word, available late at night for emotional support, genuinely cares about students' futures. The rebuttal community consensus is very strongly in his favor.
🏛 Chinese Academy of Sciences
Dingwen Tao
Extremely high pressure with sub-one-year dismissals. Changed institutions three times in six years, returning to China with very late notice to students — and continued recruiting new students to his old position after deciding to leave.
Extremely demanding — students who don't produce enough are dismissed in under a year.
Jumped between 3 universities in the US over 6 years, then returned to China. When returning, gave very late notice to students and continued recruiting new students to IU even after deciding to leave.
Described as motivated purely by career advancement/power — all students are potential "expendables."
🏛 Caltech
Anima Anandkumar
Demanding and volatile but resource-rich. Reportedly dismissed from NVIDIA after internal complaints. Runs large intern cohorts with no guarantee of support — poor performers are cut. Strong outcomes for students who survive, including top-10 faculty positions.
Reportedly requires prior collaboration before agreeing to advise PhD students.
Reported to have been dismissed from NVIDIA following complaints from lab members; joined Caltech afterward.
Recruits dozens of interns with a "survival of the fittest" approach — anyone not performing well can be let go at any time.
Reportedly sometimes emotionally unstable.
Mixed assessment
She is described as extremely attentive to certain projects (NO-related work especially), resourceful with excellent industry connections (Nvidia, top-4 labs). Lab student outcomes are reportedly very good — tenure-track at top-10, top-tier industry jobs. However, she's confirmed to be a demanding advisor who pushes hard. One note: "You need a strong heart to work with her — but resources are real and ample."
Yang Song
Accepted PhD students for Caltech before joining, then repeatedly deferred his start and ultimately went full-time at OpenAI — leaving students without a supervisor in a CS-limited environment.
Accepted students before formally joining Caltech, then deferred his start date repeatedly — ultimately never joined (went full-time at OpenAI). Students were left to find other supervisors at Caltech, which has limited CS/AI faculty.
🏛 University of Virginia
Xiaoxuan Yang
Exhibits NPD-type behaviour: publicly says 'don't rush' while privately furious at any week without progress. Insufficient funding for number of students recruited. Badmouths students to colleagues. Provides no substantive research guidance — only vague demands for 'unique innovation'.
Appears to exhibit NPD-type personality; classic "gaslighting" behavior.
Reportedly has funding for only 1–2 students but accepted 3 — leaving someone without support.
Says "don't rush" publicly but pushes intensely in private — if there's no progress within a week, reportedly livid.
Encourages internal lab competition (gladiator pit).
Reportedly badmouths students to the department chair and her own former advisor.
Multiple meetings per week plus mandatory weekly report sheets (with illogical structure).
Provides no useful research guidance — only chants "find a unique innovation point" without specifics.
Trained under Yiran Chen (see that entry for more context).
🏛 George Washington University
Howie Huang
During his promotion years recruited 5 PhD students then dismissed 3 within months of arrival, claiming he couldn't identify their strengths. Required dismissed students to stay on as OPT workers.
Described as a "gladiator specialist" — during his associate-to-full-professor years, reportedly recruited 5 students then dismissed 3. Within 3 months he would claim not to know a student's strengths, then require the student to work as an OPT employee. Described as "disgusting."
🏛 NYU
Sherry Yang
Maintains an oversized candidate pool and is reportedly arrogant in interviews. Very limited information — no specific advisor-student relationship complaints documented.
Described as maintaining a very large pool of interview candidates — and very arrogant during interviews.
🏛 Lehigh University
Lichao Sun
Primarily a research quality controversy: accused of mass-producing surveys and hype-driven papers, manipulating peer review, and running high-pressure late-night lab culture. Lab members and interns strongly dispute mistreatment, reporting good mentorship, resources, and career support.
Criticism of research quality: accused of producing large volumes of surveys and trendy papers (e.g., a "Sora" survey, "Mora," "Bora"); capable of completing a survey in 2 weeks. Accused of naming papers to ride hype cycles.
One claim that a Nature-family paper's data was fabricated — disputed by others who say the code/checkpoints are fully open-sourced and reproducible.
Posts on Xiaohongshu about career advice (e.g., "you don't need a PhD") then reverses position days later.
Accused of manipulating the review process — caught by senior researchers.
Accused of "gladiator" midnight meetings and pushing students to compete.
Strong rebuttal from students and interns
PhD students (5–6 in the lab) and many interns all report positive experiences: given freedom, resources, and connections. He is described as helping with internship placements, writing recommendation letters, and taking ~2-hour personal phone calls with students about their career development. Multiple collaborators vouch for his integrity. The research criticism (surveys, etc.) is acknowledged but countered as not being an ethical violation. A path to EB1A/NIW is mentioned as a practical benefit.
🏛 University of Minnesota
Katie Zhao
Described as very demanding; one student quit. Directly rebutted by someone who says she is easygoing and flexible with deadlines. Low-severity, disputed.
Described as very demanding; one student reportedly quit.
Rebuttal
One person says Zhao is actually easygoing and doesn't impose hard deadlines; if a deadline is missed, you just aim for the next one. Only one quit is confirmed.
🏛 University of Pittsburgh
Junyu Liu
Serious and specific sexual harassment allegations: repeatedly invites women to share hotel rooms at conferences after rejection; makes sexual advances on becoming familiar; has made openly sexist remarks about women in academia.
Warning to women specifically: Accused of widespread sexual harassment of women in academia (and outside) — reportedly repeatedly invites women to share his hotel room at conferences, and continues after rejection.
Uses a friendly public persona to get close to people, then begins making sexual advances once familiar.
Has reportedly made sexist statements privately, such as "Female PhD students all end up as prostitutes."