系統設計者的話：學英文根本不用背，你講中文，你有在背中文嗎？學英文也是一樣的道理，那是自然而然就會的東西，用背的方式最後肯定都是不行的。 更多...

最好的學習就是有英文的父母，不斷的跟你說話、耐心的教導你，時間長了，你也就慢慢就融入其中了。

但是我們沒有英文父母怎麼辦？看影片學英文就學的起來嗎？學不起來。

我們從小到大，也看過了不少英文電影，至少也看過1、200部也跑不掉，但是為什麼看了這麼多的英文電影，英文還是很爛呢？

主要的原因是我們沒有從基礎性的強化學習。舉個例子，西方的幼兒出生時，他至少聽了各種大人講英文至少聽了2年，才慢慢開始說英文、與大人互動。

而且都是一字一句慢慢的說、慢慢的教、漸進式的、慢慢的、越來越習慣、越來越聽的懂、越來越會說的過程。

那麼我們看的影片、電影過程中，都是大人成熟型的講話速度、方式、語調，我們怎麼可能跟的上？

一旦跟不上，思想中就會產生覺的好難哦、我不行種種念頭，那麼產生了這種潛在式的排斥心理作用，那麼你就更難學的會了。


所以我們必須打好基礎、深厚的基礎，萬丈高樓平地起，唯有打下深厚的基礎，後面你走的路才會越來越順。

程式界裏有句話，程式債欠的越多，最後就只能放棄。學英文也是一樣，英文單字債欠的越多，你就直接上來文法、句型、語調，最後也只能放棄。

基於上述的思維，我打造了這套系統，從最基礎的單字學習起，不斷的強化、加深你對這個單字的聽力，透過各式各樣的影片、各式各樣的人、各種不同的場景。

就只是為了訓練一個單字的聽力，無數的人跟你講同一個單字，講個20遍、50篇、100遍(且不重覆)，我就不信學不起來。


這套系統做出來後，我自己在使用的過程中，確實真的能聽出來他們講的這個單字，而且他們講的速度再快也能聽的出來。

能聽的出來，就知道他講的大概意思；能夠知道他講的大概意思，就能夠用幾個簡單的單字回應；能夠回應，也就能夠對話；當能夠對話，就能夠再修正文法、句型、語調。


我的目地就是希望自己至少能聽的懂人家在說什麼，用英文跟人家對話，大家共同努力、學習。若有其它想法或對這套系統設計的建議，也可以與我聯絡：ichich2013@gmail.com

Keynote (TensorFlow Dev Summit 2018)
觀看次數：340
i'm
anita
vijay
kumar
technical
program
manager
for
tensorflow
and
welcome
to
the
2018
tensorflow
transformative
journey
in
the
field
of
this
field
is
unleashing
new
frontiers
we've
already
seen
great
advantages
in
star
researchers
and
astrophysicists
risk
for
cardiovascular
diseases
just
affordable
non-invasive
screening
for
ensuring
that
era
aircrafts
can
fly
through
the
air
spaces
of
belgium
luxembourg
germany
and
the
Netherlands
built
real-time
detection
and
alerting
mechanisms
to
detect
logging
of
trees
deforestation
and
protecting
our
deep
neural
networks
that
can
learn
the
characteristics
of
sounds
and
create
characteristics
in
arts
and
culture
you
every
product
is
being
fundamentally
these
examples
tensorflow
is
at
the
forefront
of
them
tensorflow
powerful
scalable
and
the
with
lots
of
new
releases
if
you've
seen
we've
focused
on
making
tensorflow
high-level
api's
you
know
like
tf
data
in
tf
carris
and
new
programming
self-identified
locations
of
folks
on
github
that
have
starred
tensorflow
and
antarctica
in
fact
tensorflow
is
used
in
important
part
of
any
open-source
questions
responding
to
queries
and
so
launching
a
new
tensorflow
blog
will
be
of
conversations
and
interviews
with
the
speakers
to
make
reuse
and
sharing
easier
today
we
are
launching
tensorflow
focused
on
introducing
this
to
a
much
deep
learning
and
neural
networks
since
algorithms
like
linear
regressions
and
state-of-the-art
tools
for
stats
and
make
tensorflow
easy
to
use
zarnow
easier
to
use
tensorflow
also
pre-made
estimators
a
library
of
ready
to
go
implementations
of
common
machine
learning
algorithms
so
let's
take
a
look
computing
gradients
and
updating
the
use
your
python
debugger
to
debug
that
have
at
kara's
in
this
case
it
executes
use
more
than
one
accelerator
or
gpu
or
devices
that's
lightweight
and
it's
integrate
tensorflow
lite
with
a
few
don't
miss
the
talk
by
Chris
lackner
platform
javascript
is
a
language
that's
flora
jas
it's
a
it's
bringing
machine
full-fledged
layers
API
on
top
it
also
covers
these
api's
are
accelerated
in
the
browser
via
webgl
and
then
the
tensorflow
tax
rate
on
cpus
and
gpus
as
we're
measuring
and
assessing
intent
important
to
have
reproducible
sure
that
tensorflow
performs
well
highlighting
tensorflow
light
speed-up
when
using
tensorflow
like
quantized
versus
standard
tensorflow
that
come
out
of
the
collaborations
that
tensorflow
performs
well
on
these
so
there's
the
Intel
mkl
dnn
library
optimized
for
tensorflow
strong
collaborations
with
different
collaborations
we've
had
with
nvidia
as
we're
excited
about
is
tensor
rt
sensor
runtime
for
gpu
platforms
we've
been
collaborating
very
closely
with
nvidia
on
tensor
rt
10
so
our
t's
been
around
tensorflow
17
release
we
now
have
making
use
of
tensor
rt
standard
tensorflow
so
it's
great
again
to
see
these
types
of
collaborations
that's
how
you'll
get
the
best
out
of
tesla
v
100
that
nvidia
has
so
we've
looking
at
examples
for
real-world
data
benchmark
on
synthetic
data
but
it's
this
is
google's
v2
tpu
so
it's
designed
computation
but
it's
not
just
about
the
that
we've
already
open-source
that
are
about
cloud
gpus
is
the
availability
of
year
so
what's
a
pod
a
pod
is
actually
sixty-four
of
the
devices
like
I
showed
get
about
115
petaflop
of
computation
in
a
pod
so
that's
a
lot
we're
focusing
on
i've
given
you
some
understanding
of
chemical
molecules
and
someone
who's
downloaded
tensorflow
and
tens
of
thousands
of
organizations
in
optimize
a
controller
that
proposes
don't
construct
but
it
has
characteristics
of
computational
layers
so
the
interesting
computational
budget
but
it's
also
true
look
very
low
computational
cost
and
categories
of
problems
okay
in
ophthalmology
where
you're
trying
to
diabetic
retinopathy
this
is
a
serious
degenerative
eye
disease
400
million
world
and
it's
very
treatable
if
it's
ophthalmologists
to
inspect
these
images
was
on
par
with
board-certified
ophthalmologists
and
since
then
we've
ophthalmologists
and
then
we
actually
but
more
interestingly
i'm
going
to
tell
the
the
retinopathy
team
and
as
a
warm-up
exercise
Lily
panghu
that's
weird
cardiovascular
health
you
know
your
age
cardiovascular
health
as
are
things
like
non-invasive
test
for
cardiovascular
been
collaborating
with
several
healthcare
organizations
to
work
on
de-identified
medical
records
to
see
if
predicting
which
patients
are
most
at
risk
for
mortality
and
using
this
we're
good
examples
of
where
there's
real
affect
cassava
and
these
diseases
makes
use
to
tensorflow
images
of
different
cassavetes
diseases
when
lm
abundantly
onna
when
you
love

相關影片

加Line好友 每日8AM 推播1單字

加入好友

其他人正在學習的單字

1. just
2024-05-03 07:33
2. board
2024-05-03 07:33
3. animate
2024-05-03 07:33
4. border
2024-05-03 07:33
5. freeto
2024-05-03 07:33
6. daughter
2024-05-03 07:33
7. ordinary
2024-05-03 07:33
8. ability
2024-05-03 07:33
9. exception
2024-05-03 07:32
10. ad
2024-05-03 07:32
11. um
2024-05-03 07:32
12. pieces
2024-05-03 07:32
13. car
2024-05-03 07:32
14. sell
2024-05-03 07:32
15. powerful
2024-05-03 07:32

本月觀看影片單字排行榜

本月查閱單字排行榜

本月聆聽單字排行榜