dismissed EB-1A Case: Artificial Intelligence

U.S. Citizenship 
and Immigration 
Services 
In Re : 10768211 
Appeal of Nebraska Service Center Decision 
Non-Precedent Decision of the 
Administrative Appeals Office 
Date : OCT . 5, 2020 
Form 1-140, Immigrant Petition for Alien Worker (Extraordinary Ability) 
The Petitioner, a provider of financial and business data and news, seeks to classify the Beneficiary as 
an alien of extraordinary ability in the field of artificial intelligence (AI). See Immigration and Nationality 
Act (the Act) section 203(b)(l)(A) , 8 U.S.C . § 1153(b)(l)(A) . This first preference classification makes 
immigrant visas available to those who can demonstrate their extraordinary ability through sustained 
national or international acclaim and whose achievements have been recognized in their field through 
extensive documentation. 
The Director of the Nebraska Service Center denied the petition, concluding that while the Beneficiary 
met the initial evidence requirements of the requested classification , the record did not establish that 
he has sustained national or international acclaim and is among the small percentage at the top of his 
field. On appeal, the Petitioner asserts that the Director imposed novel requirements in rendering his 
decision, and that the Beneficiary qualifies as an alien of extraordinary ability . 
In these proceedings , it is the Petitioner's burden to establish eligibility for the requested benefit. See 
Section 291 of the Act, 8 U.S .C. § 1361. Upon de nova review , we will dismiss the appeal. 
I. LAW 
Section 203(b)(l) of the Act makes visas available to immigrants with extraordinary ability if: 
(i) the alien has extraordinary ability in the sciences, arts, education, business, or 
athletics which has been demonstrated by sustained national or international 
acclaim and whose achievements have been recognized in the field through 
extensive documentation, 
(ii) the alien seeks to enter the United States to continue work in the area of 
extraordinary ability, and 
(iii) the alien's entry into the United States will substantially benefit prospectively 
the United States . 
The term "extraordinary ability" refers only to those individuals in "that small percentage who have 
risen to the very top of the field of endeavor." 8 C.F.R. § 204.5(h)(2). The implementing regulation 
at 8 C.F.R. § 204.5(h)(3) sets forth a multi-part analysis. First, a petitioner can demonstrate 
international recognition of his or her achievements in the field through a one-time achievement (that 
is, a major, internationally recognized award). If that petitioner does not submit this evidence, then 
he or she must provide sufficient qualifying documentation that meets at least three of the ten criteria 
listed at 8 C.F.R. § 204.5(h)(3)(i) - (x) (including items such as awards, published material in certain 
media, and scholarly articles). The regulation at 8 C.F.R. § 204.5(h)(4) allows a petitioner to submit 
comparable material if they are able to demonstrate that the standards at 8 C.F.R. § 204.5(h)(3)(i)-(x) 
do not readily apply to the individual's occupation. 
Where a petitioner meets these initial evidence requirements, we then consider the totality of the 
material provided in a final merits determination and assess whether the record shows sustained 
national or international acclaim and demonstrates that the individual is among the small percentage 
at the very top of the field of endeavor. See Kazarian v. USCIS, 596 F.3d 1115 (9th Cir. 2010) 
( discussing a two-part review where the documentation is first counted and then, if fulfilling the 
required number of criteria, considered in the context of a final merits determination); see also 
Visinscaia v. Beers, 4 F. Supp. 3d 126, 131-32 (D.D.C. 2013); Rijal v. USCIS, 772 F. Supp. 2d 1339 
(W.D. Wash. 2011). 
II. ANALYSIS 
The Beneficiary is currently employed by the Petitioner as a senior research scientist in its Artificial 
Intelligence Group, where he conducts research on natural la~rocessing (NLP) and machine 
learning (ML). He earned a Ph.D. in computer science froml___JUniversity in 2017, and states 
that he intends to continue working in this area in the United States. 
A. Evidentiary Criteria 
Because the Petitioner has not indicated or established that the Beneficiary has received a major, 
internationally recognized award, he must satisfy at least three of the alternate regulatory criteria at 
8 C.F.R. § 204.5(h)(3)(i)-(x). The Director found that the Beneficiary met three of the evidentiary 
criteria at 8 C.F.R. § 204.5(h)(3)(i)-(x), relating to his participation as a judge of the work of others, 
his authorship of scholarly articles, and his receipt of a salary that is high in comparison to others in 
his field. On appeal, the Petitioner asserts that the Beneficiary also meets two additional evidentiary 
criteria. 1 After reviewing all of the evidence in the record, we agree that the Beneficiary meets the 
three criteria referred to above. As the Beneficiary has therefore met the initial evidence requirement, 
we will not consider whether he also meets the additional claimed criteria, but will instead consider 
all of the evidence in the record in performing a final merits determination. 
1 The Petitioner also claims to have submitted comparable evidence under 8 C.F.R. § 204.5(h)(4), but does not explain 
why the criteria listed under 8 C.F.R. § 204.5(h)(3) are not readily applicable to the Beneficiary's occupation. Further, as 
we agree with the Director's finding that the Beneficiary meets the initial evidentiary requirements for the requested 
classification, the Petitioner's claim of comparable evidence is moot and thus will not be considered. 
2 
B. Final Merits Determination 
The record indicates that the Beneficiary's work in AI has been concentrated in the sub-fields ofNLP 
and ML, and that he continues to focus his work in these areas with the Petitioner. On appeal, the 
Petitioner asserts that the Beneficiary has made three main contributions to these areas during the 
cour~e of his studies and in its employ. The first involves his participation in the development of an 
NLPl ] which took place durin his em loyment as a research intern. In a reference letter 
submitted with the initial filing, Chief Scientist atl I describes his 
collaboration with the Beneficiary to develop '~----~ learning architectures that could be 
applied to a wide range of complex NLP problems ... " and notes that this research resulted in.---a-=-===r 
I I that could be used for various different tasks in natural lan9 u.::ca::=.::....i:..::...::....::.===..:..:........i.---___,, 
adds that this research was resented at two conferences, the 2015 
I I and the 2016 ,..::...._....:..........----==============;-' 
. However, ~----' does not provide specific examples of the application of this 
"F=====;lb_y_o~thers to solve NLP problems or perform tasks. 
The Petitioner asserts on appeal that the paper presented atl I in 2016, I ~ 
I I" "remains among the publication's top 
30 most cited articles." Although the evidence referenced by the Petitioner indicates that this paper 
was cited on 597 occasions 2 by other researchers as of the date the petition was filed, and that it 
substantially exceeds! ~s "h5-median" score, it does not support the assertion that it is among 
I ts top 30 most cited articles. 3 It is by far the most highly cited of the Beneficiary's published 
papers, and this figure by itself indicates that many researchers in the field of AI have used this paper 
in support of their own research in some way. But the record does not include evidence which 
demonstrates whether or how other researchers have applied or built upon the Beneficiary's work, or 
that the framework has been implemented by others in the field, and thus does not show the nature or 
degree of impact this work has had on the field of AI, or the level of recognition the Beneficiary has 
received because of it. 
In addition, the Petitioner also refers to evidence of the importance of the I I conference in the 
field of AI. This includes a list from the Google Scholar website showing that it ranks second in the 
field of AI in terms of "h5-index" and "h5-median," terms which reflect that a certain number of 
articles published by that medium have been cited the same number of times in the past five years. 4 
In addition,! I states in his letter thatl lis "one of the primary conferences of high impact 
in Machine Leaming and Artificial Intelligence research." However, the Petitioner does not explain 
how this ranking affects the impact or recognition of the Beneficiary's work beyond the factors 
discussed above. We will not assume that a paper published in a prestigious or highly-ranked journal 
or conference proceeding is inherently more impactfol or gains greater recognition than a paper 
published in a different medium. 
2 This paper had 909 citations as of September 1, 2020 per the Beneficiary's Google Scholar profile, 
https:/l I 
3 We also note that since the h5-median reflects citations to papers published in the period from 2015-19, it is not an 
accurate gauge of the impact of the Beneficiary's work compared to more recently published articles which have had a 
shorter period in which to be cited, nor does it take into account citations to papers published in this journal before 2015. 
4 The Petitioner also refers to evidence that this conference (and others at which the Beneficiary presented his work) has a 
"CO RE rank" of A*, but the record does not include evidence that explains either the meaning or significance of this rank. 
3 
The second contribution highlighted by the Petitioner is based upon the Beneficiary's work published 
at the I I conference in 2014, '.__ _____________________ ____. 
I I" As with the paper discussed above, the Petitioner focuses on the number of citations to 
this paper, 193 as of the time of filing, and points out that this is above the conference proceeding's 
h5-index (but below the h5-median figure.) While we acknowledge that this citation figure indicates 
that many other researchers in the field of AI have relied on the Beneficiary's work in some way, we 
note that this evidence does not provide sufficient qualitative or quantitative context to establish that 
the he is one of the small percentage at the top of this field. Although the h5-index fo!LJ indicates 
that the paper is one of the more highly-cited papers to be presented at the conference in the past five 
years, it does not provide information regarding papers published in the field overall at other 
conferences and in scientific journals, or papers which were previously published in the I I 
proceedings. In addition, it does not provide insight into how or to what extent the researchers who 
cited the Beneficiary's paper used this research in their own work. 
Reference letters discussing this work provide some farther insight into its impact on the field of AI. 
~-------~ ot1 I University, who collaborated with the Beneficiary on this work, 
confirms that the Beneficiary conducted the bulk of the work towards this research under her guidance, 
and concludes that it "has had a far-reaching impact on the worldwide natural language community ... " 
But she does not provide details as to the nature of this impact, or provide specific examples of how it 
has been recognized by others in the AI field. 
Another letter was submitted by~-------~ofl !University, who indicates 
that he does not know the Beneficiary but is familiar with his work. He describes the Beneficiary's 
research "on I I interpretation for ~-----~ in natural language processing" as 
"revolutionary," and states that he developed a new method which works in conjunction with previous 
work but "overcame some of their drawbacks." I !indicates that this work "provides 
insight about different! I components that helps the industry design better architectures for 
new tasks," and is relevant to a company such as the Petitioner, but he does not provide details about 
whether this model has been implemented in the industry or otherwise recognized. 
The third contribution which the Petitioner notes on a eal relates to his work which was published at 
,.......... ......... ..__ _________________ ,___---=----.,...J conference in 2014, I I 
As with the previous examples, the Petitioner 
compares the number of citations to this paper with the publication's h5-index, noting that its 236 
citations triples that figure, and includes evidence showing that the conference proceedings are ranked 
second among publications in the field ofl I Again, we acknowledge that 
these figures indicate that the Beneficiary's work has had impact on the AI field, but note that they do 
not provide a complete picture as to the extent and nature of that impact. 
More information regarding this aspect of the Beneficiary's work is provided by~I -----~ 
I I of the University! I who writes in his reference letter that he first met the 
Beneficiary at this conference. He states that being selected to present at this conference "is a 
tremendous honor that separates [the Beneficiary] from others in his field," but he does not provide 
figures such as the rate of acceptance of submitted a ers or the total number of papers presented at 
the conference, to support this assertion. oes on to state that the Beneficiary's 
paper "is by far the best research on the task of........,. ___ ~--~extraction," and notes that "this 
4 
was the first application of~-------~to anyl I task." He farther indicates that 
this research "is of practical importance" since it removes the 1 I engineering" step and thus 
saves time and money. However, he does not explain the extent to which the Beneficiary's work has 
affected that of other AI researchers, been implemented by them, or been acclaimed in the industry in 
other ways. 
As we have noted in the above analysis, while some of the reference letters submitted provide detail 
regarding the technical aspects of the Beneficiary's work, none provide the same level of specificity 
when explaining its impact upon the field of AI. Depending on the specificity, detail, and credibility 
of a letter, USCIS may give the document more or less persuasive weight in a proceeding. The Board 
of Immigration Appeals (the Board) has held that testimony should not be disregarded simply because 
it is "self-serving." See, e.g., Matter of S-A-, 22 I&N Dec. 1328, 1332 (BIA 2000) (citing cases). The 
Board also held, however: "We not only encourage, but require the introduction of corroborative 
testimonial and documentary evidence, where available." Id. If testimonial evidence lacks specificity, 
detail, or credibility, there is a greater need for the petitioner to submit corroborative evidence. Matter 
of Y-B-, 21 I&N Dec. 1136 (BIA 1998). Here, several of the letter writers conclude that the 
Beneficiary's work is "relevant" to the AI field and industry, but do not provide farther details about 
how it has gone beyond relevance and gained sustained acclaim at the national or international level. 
The conclusions that several of the writers make regarding his elevated standing in the field are 
therefore not supported by their own statements, nor are they corroborated by documentary evidence. 
The Petitioner also submitted evidence regarding the Beneficia['s work since joining its research 
team in 2017, as well as the role he has played for the company. I Head of AI 
Engineering for the Petitioner, writes that the Beneficiary has continued his research inl I 
I I for text analysis, and that models he has developed have improved the performance of 
software "by nearly 5% in precision and recall." He farther states that these models and software are 
used by several of the Petitioner's teams and products. In addition,! lnotes that the 
Beneficiary has co-instructed four courses in ML and NLP given to the Petitioner's engineers, which 
has led to the development of four machine learning products within the company. Although the 
Beneficiary has clearly contributed to the Petitioner's research efforts in the field of AI, and his work 
has improved the efficiency of the work of others employed by the Petitioner, the evidence does not 
establish that he has played a critical role for the company's overall organization. Although the 
systems and products improved through the Beneficiary's efforts are important to the Petitioner's core 
business, the evidence does not show that the incremental upgrades accomplished through the 
application of his work have been critical to the company's overall success. 
In addition, this reference letter does not establish that the Beneficiary's work during his employment 
with the Petitioner has received acclaim in the broader AI field. I I writes that the results 
,.=J.:;.u;:.....1.1..1.ll.ll.u...i.u:.ul....'1.L...llJJ;:;....<:.l..L..J....:z!:=-=-...,..=.,..,,,.....,.........,,..,...,...,...,.,.....,...,...,..,........,.,.....,,,,.,.....,...,~, part of the......., _______ .....,....... 
~------------- conference, "the leading venue for researchers around the world to 
publish,__ ________ related ideas." We acknowledge that the dissemination of this work 
makes it available to researchers in the AI field beyond those employed by the Petitioner, but the 
record does not include evidence which shows that it garnered national or international acclaim for the 
Beneficiary. 
5 
Regarding the Beneficiary's participation as a peer reviewer, the Petitioner argues on appeal that the 
Director employed novel legal standards in finding that this participation was routine and did not 
exceed that of other researchers, and that he ignored the evidence of the acclaim of those journals and 
conferences. As for the first point, we note that the Director did not make these findings when 
determining whether the Beneficiary met the plain language requirements of the criterion at 
8 C.F.R. § 204.5(h)(3)(iv), but when considering whether he has garnered sustained national or 
international acclaim and is one of the small percentage at the top of his field. This analysis necessarily 
involves a comparison of the Beneficiary's accomplishments against those of all other researchers in 
his field, and we conclude that the Director's findings are pertinent and well within the framework of 
the final merits determination per Kazarian. 
Turning to the Petitioner's second point about the Beneficiary's work as a peer reviewer, it notes that 
he conducted this work for some of the same highly-ranked conferences at which his work was 
published, includin~ I, D and I I The Petitioner states that the Beneficiary "has not 
simply participated in widespread peer review processes. He has conducted peer reviews for the 
largest, well-known and prestigious conferences and publications in his field of endeavor." While we 
acknowledge that scientific conferences and journals rely upon experts such as the Beneficiary to 
review papers which are submitted for presentation or publication, the Petitioner has not explained 
how the prestige of these conferences bestows acclaim on the many anonymous reviewers that are 
needed to provide recommendations to session chairs making the ultimate decision on which papers 
will be presented. 
The record also includes evidence of the Beneficiary's salary, as well as comparative data showing 
that it exceeds the national 90th percentile salary for the position of "Computer and Information 
Research Scientists." We agree that this evidence shows that his salary is high when compared to 
others in his field, and reflects the level of education and expertise he possesses. However, when 
considered together with the totality of the evidence as analyzed above, this evidence does not 
sufficiently demonstrate that the Beneficiary is one of the small percentage at the top if his field. 
III. CONCLUSION 
The evidence shows that the Beneficiary is an accomplished researcher who has already made some 
degree of impact in the field of artificial intelligence. Although he has published several articles in 
the proceedings of scientific conferences, some of which have been cited by other researchers 
hundreds of times, the record does not contain evidence which establishes that this work has brought 
him acclaim at the national or international level. In addition, his publication record, participation in 
the peer review process, and salary, when considered together, do not show that he is among the small 
percentage of AI researchers at the top of the field. 
The Petitioner seeks a highly restrictive visa classification, intended for individuals already at the top 
of their respective fields, rather than for individuals progressing toward the top. USCIS has long held 
that even athletes performing at the major league level do not automatically meet the "extraordinary 
ability" standard. Matter of Price, 20 I&N Dec. 953,954 (Assoc. Comm'r 1994). Here, the Petitioner 
has not shown that the significance of the Beneficiary's work is indicative of the required sustained 
national or international acclaim or that it is consistent with a "career of acclaimed work in the field" 
6 
as contemplated by Congress. H.R. Rep. No. 101-723, 59 (Sept. 19, 1990); see also section 
203(b)(l)(A) of the Act. 
For the reasons discussed above, the Petitioner has not demonstrated the Beneficiary's eligibility as 
an individual of extraordinary ability. The appeal will be dismissed for the above stated reasons, with 
each considered as an independent and alternate basis for the decision. 
ORDER: The appeal is dismissed. 
7